development

[cyan niche]

rn I work for the government as a contractor

Previously I was doing code consulting for some large companies.

[sly marten]

I’m basically IT support for $60k a year, so I’d say it’s pretty good considering for me, enjoyable work too, chill, not so stressful and painful

[cyan niche]

¯_(ツ)_/¯

Never merge on Fridays and I don't feel any stress xD

[sly marten]

It gives me time and energy to do other stuff I want to do on the side

Even during work hours

[cyan niche]

My real tip to other devs is get a remote job, then never actually put in eight hours a day, put in like 4-6 most days.

[sly marten]

I tried working remote, not for me, I get too easily distracted

[cyan niche]

You're only ever putting in that much anyway, and it frees up your time to decompress without being stressed about it.

Oooohhh, yeah remote work is the life.

[sly marten]

But yeah, it’s a great tip if you can manage a remote job

No employer will bat an eye as long as you’re not obviously slacking

[cyan niche]

Who knows if I'm slacking or not? xD

[sly marten]

Project deadlines and estimates 😛

[cyan niche]

Pff, they base those off of my output.

[sly marten]

That’s what killed me, not because I was slacking, but because I kept pressuring myself, so I was my problem, nothing else really 😂

[sly marten]

My estimates were based off of estimates output of an average developer, or whatever someone felt like we’re good estimates anyway

[peak acorn]

I'm working remote as intern until I graduate I'll tell u my productivity absolutely tanked from when I was going in to the office over the summer lmao

They still pay me tho so idc!!

[cyan niche]

Yeah I can do much more than the average developer, I just choose to hit around the average.

[sly marten]

Well they want consistency more than bursts of super productivity probably anyway, and if working a little less productive means consistency, they’ll be happy, and so will you probably

[sly marten]

Yeah, I just don’t really have it in me to slack, I keep pushing 😂

Probably a lot because of my ADHD

[cyan niche]

Yeah, companies will just not care, they don't give you big raises for it, and just keep piling more on your plate.

That's when I said fuck it dude I'm going to work like two days a week.

[sly marten]

So development will be particularly exhausting I guess, because it really never ends, and the more you output, the more they’ll throw at you

Yeah, i would probably make a significantly higher salary by now if I had managed to keep that phase 😛 but it’s not sustainable, especially when you’re also a perfectionist and you’re assigned to work on things that feel kinda pointless a lot of the time

Such as getting pictures from some random designers on how a website is supposed to look, then you are to implement that design in a system where you’re very limited on how much HTML you can change, and also unable to remove underlying CSS, so you kinda have to pile shit CSS on top just so it’ll work (such as using !important for everything)

[cyan niche]

Ask for mo money

:D

[sly marten]

I was burnt out long before I got to that point, also when I eventually started burning out, I felt so useless, and it ended up in a very bad circle lol

So never found a way to justify more pay without first getting my ass out of the gutter

Health system also failed me big time, didn’t get sick leave, but also didn’t get help, but clearly was struggling with my health, even my doc knew 😂

[cyan niche]

dayuum

[sly marten]

Very unfortunate, but it is what it is, and I’m back on track now 😎

[cyan niche]

gotta look after #1

[sly marten]

ADHD meds changed everything quite a bit, but so did a different type of job too

Now I can develop what I want to develop, when I want to develop it, no expectations or anything

[cyan niche]

nono money is the purpose of life not happiness

[sly marten]

I’m on the path of making more money now, I even make more in my current role that I did as a developer 😂

I did take on this task at work to create this web app that needs to work in a IE11 browser though 😂 and I’ve over complicated it (probably), as usual haha

So I’m really pressed for time if I want to meet the wished deadline, though I’ve never promised anything when it comes to a deadline

But it will be miles better than what we currently have at least, much more maintainable and easy to work with too

[deep flicker]

Oh I never claimed to understand it. That code is some mystical magical voodoo. Given the number of sin and cos terms, I think it's some sort of spherical coordinate mapping into an ASCII array. If you run the program at a low frame rate it looks like it has some sort of lighting/shadows worked into the ASCII array.

for(j=0; 6.28>j; j+=0.07) 
{ 
  for(i=0; 6.28 >i; i+=0.02) 
  { 
    float sini=sin(i), cosj=cos(j), sinA=sin(A), sinj=sin(j), cosA=cos(A), cosj2=cosj+2, mess=1/(sini*cosj2*sinA+sinj*cosA+5), cosi=cos(i), cosB=cos(B), sinB=sin(B), t=sini*cosj2*cosA-sinj* sinA; 
    int x=40+30*mess*(cosi*cosj2*cosB-t*sinB), y= 12+15*mess*(cosi*cosj2*sinB +t*cosB), o=x+80*y, N=8*((sinj*sinA-sini*cosj*cosA)*cosB-sini*cosj*sinA-sinj*cosA-cosi *cosj*sinB); 
    if(22>y&&y>0&&x>0&&80>x&&mess>z[o])
    { 
      z[o]=mess; 
      b[o]=".,-~:;=!*#$@"[N>0?N:0]; 
    } 
  } 
}

Truly understanding that program probably requires a ritual and some sort of animal sacrifice 😛

[peak acorn]

Reverse engineering that sounds horrible

[spring cradle]

oh cool

[twilit beacon]

this is even worse

k;double sin() ,cos();main(){float A= 0,B=0,i,j,z[1760];char b[ 1760];printf("\x1b[2J");for(;; ){memset(b,32,1760);memset(z,0,7040) ;for(j=0;6.28>j;j+=0.07)for(i=0;6.28 >i;i+=0.02){float c=sin(i),d=cos(j),e= sin(A),f=sin(j),g=cos(A),h=d+2,D=1/(c* h*e+f*g+5),l=cos (i),m=cos(B),n=s\ in(B),t=c*h*g-f* e;int x=40+30*D* (l*h*m-t*n),y= 12+15*D*(l*h*n +t*m),o=x+80*y, N=8*((f*e-c*d*g )*m-c*d*e-f*g-l *d*n);if(22>y&& y>0&&x>0&&80>x&&D>z[o]){z[o]=D;;;b[o]= ".,-~:;=!*#$@"[N>0?N:0];}}/*#****!!-*/ printf("\x1b[H");for(k=0;1761>k;k++) putchar(k%80?b[k]:10);A+=0.04;B+= 0.02;}}/*****####*******!!=;:~ ~::==!!!**********!!!==::- .,~~;;;========;;;:~-. ..,--------,*/ 

[twilit beacon]

just as i thoight the donut code was crazy i foind this videom on tiktok

if its legit, i cant even comprehend how that would be in any way possible

[hollow basalt]

From the video it's 8180 which has 56 threads.
It uses the 32 sockets, so 32*56 = 1792 (seen from the video)

I didn't bother checking how they did that.
But seems possible to simulate doom in that scenario

[peak acorn]

Can you somehow target specific cores?

[midnight wind]

I think they just hooked into the task manager

Not actually making the usage change

It refreshes way too fast

[peak acorn]

True

[silk eagle]

god I love Bunny's CDN so much

[peak acorn]

I need a cdn dedicated to bunny pictures

[north raft]

Am i allowed to ask for someone to do some plugin work for my server here? rules were a little unclear of this part

[trail remnant]

This reminds me of the International Obfuscated C Code Contest https://www.ioccc.org/years.html#1998

This is the C file that won the "best in show" award in 1998.
https://www.ioccc.org/1998/banks.c

It's a flight simulator that looks like a plane.

[peak acorn]

I believe the animated donut was an ioccc submission

[spring cradle]

does that preserve compilability?

[slate frigate]

If any LMG engineers ever frequent this channel, I'm SUPER interested in how y'all are handling the interactions between the golang gui and python framework with the new "Mark" software

Have an app that I'm looking to add easy user plugin creation, but still haven't settled on a way I'm entirely satisfied with

[spring cradle]

that is to say, does the donutized source still compile

[peak acorn]

It does

its a pretty famous program

[spring cradle]

niice

[celest junco]

Yeah that seemed like a weird combo to me (go isnt exactly known for front end dev). I was also curious about:

1. The decision to have a GUI at all versus just a CLI.
2. Why convert from CSV to protobuf
3. How do they implement the actual benchmarks? I may have missed it but I thought it didnt get covered in the video and I thought that would be the most complex part.

[slate frigate]

There pulling frame data directly from gpu, and HW data directly too, and the protobufs just make it easier to send everything around. As for the mix of go and python, my guess is that the go handles the gui and DB connections, with the python is what sets up and runs the benchmarks. Not a bad approach, especially considering how finicky setup and run would be game from game, lets them write tests as individual plugins rather than worry about a bad test killing the whole program @celest junco

[celest junco]

what I understood is they have some program dumping out the frame data in CSV format... why bother converting it to protobuf?

just ship the csv

[slate frigate]

Probavly because they're sending it over the network, and if they're running multiple tests at the same time, probably better for performance and reliability. Maybe they're using grpc? Thats one of the ways I've explored writing plugins

If they are using something other than http for comms, that'd explain the protobufs to me

[celest junco]

oh one other thing... if anyone from LTT is reading: It would be interesting to see this being open sourced in the future

[slate frigate]

I think thats their plan

@celest junco and go actually has a pretty decent gui framework with Fyne. I have some complaints with how they handle images, but its pretty solid. Way better than going at glfw completely raw. I THINK thats what ltt is using based off the very limit view we had of it. Styling and layouts looked familiar

[celest junco]

any thoughts on my 3rd point?

[slate frigate]

They're writing handler scripts, probably using pygame or something similar, to start and run em

[celest junco]

I dont follow how would pygame automate playing games?

[slate frigate]

Excuse me, pyautogui, not pygame

Lets you simulate user inputs, and I've seen it used it in the past to build bots, its especially powerful when combined with a CV library

[celest junco]

yeah I thought about that, but I figured they would have to spend a very long time to get it to play games semi proficiently

Good point on using CV, they did mention they were hiring for it

I still think it would be a challenging task. Personally that would be an interesting video topic.

[slate frigate]

Nah, not if they're just running games with benchmarks. Could probably do most with just arrow buttons and enter

[twilit beacon]

arent there better ways to input things rather than pyautogui

[spare spade]

Im my opinion I agree with some of your points but ultimately it boils down to using the right tool for the job. OO had to be a thing and java does it brilliantly. That's why I love it as well lol.

[silk eagle]

I'm a python main, partly because its a dumb language for uneducated programmers, but also partly because I can get things done and ready to go within an hour or two. but to remove from the first reason, I'm currently working on learning rust which (from my perspective) is basically a cs class put into language form.

rust feels a lot more palatable to me than most of the other 'complicated' languages.

[nocturne galleon]

I'm currently a student in a software engineering bootcamp but I could really use a friend to talk to about what I'm learning and just staying focused in general

please message me if you're interested in talking about learning javascript

[hollow basalt]

ok

[cyan niche]

I missed the hot take on OOP.

But as it turns out that OOP is the natural evolution of programming from its imperative roots, and it becomes necessary very quickly to manage increasing complexity that similar things need to be grouped into a single bucket so that you, the programmer, can easily manage what's going on. Even C has structs and uses naming conventions to group like things together. A properly readable C program is already trying to do what C++ does, which is why C++ came along in the first place.

"But what about inheritance problems? Functions? Spaghetti code?"

They're all irrelevant because the limits of OOP are the programmer. OOP doesn't make you write bad code, it makes people put guard rails on their code to prevent other people, perhaps even your future self, from modifying something that shouldn't be modified, or doing something that breaks something else. Inheritance solves code duplication problems like bugs introduced from these problems. The fact that you may have to expand the parent function to make it compatible with its inheritance for one case is a feature and not a bug, as it forces the programmer to consider the scope of all related objects. And there is always the option of overloading.

In modern times, all of the arguments against what OOP languages can't do is pretty much niche and petty. Functions are first class citizens in today's OOP languages. C# has Func<> which even has support for dynamic assignment or anonymous type returns. Modern OOP is flexible enough to handle the bits of a program that must be flexible, solving whatever grievances people had with it back in the 90's. It's basically a blend of all the best parts of functional and object oriented paradigms. Even declarative typing has been worn down with var assignments. Typescript has an "any" type.

This brings me back to the point the real limitation is the developer who writes it. If your code is spaghetti code it's because you're bad, not because the language is.

Java took the OOP principles to heart, but even it has had to force to adapt to the new way of doing things. It had to add Stream API to keep up with LINQ. Honestly, it's been playing catch up to C# ever since C# 3. And Oracle has the licensing so messed up there's OpenJDK and their version and compatibility between the two can feel fucking awful. At least with .NET you get one single version that just automatically works. And don't get me started on its package managers. It also did silly things like not allow value comparisons on string comparisons because "that's the OOP way of doing things" even though no programmer would ever need to compare the references between two string objects. Other languages were just like, "Oh just override the .Equals". That's OOP too.

[cyan niche]

The hardest thing to grasp in Javascript is promises. Once you get that figured out you're home clear JS is easy.

[cyan niche]

Python is one of my favorite languages because it is really just very easy to do anything you want and mostly built in.

Best package management out of any language I've ever used too.

[spark temple]

Don’t inherit, compose.

[midnight wind]

Lol

Python package manament is a PITA

Better than C/C++, but still a mess with venvs and everything

[cyan niche]

no

that's what makes it great

[midnight wind]

Nope

[cyan niche]

It's one line virtualenv venv

[midnight wind]

Go's package management is much more pleasant to use

[cyan niche]

I've never used Go

[peak acorn]

Oop is awsom,, functional confusing!

Make me a linked list in functional lang only I'll wait freak

[spark temple]

Check this out. It's pretty good from my quick glance.

https://dev.to/joelnet/creating-a-linked-list-using-only-function-combinators-5hhk

[peak acorn]

Yeah when u need an article I think you've gone wrong

[spark temple]

And I'm not sure your background, but working in a large code base that heavily leans on OOP can be very painful when it uses inheritance all over the place.

Causes a lot of things tightly coupled when requirements begin to shift. Composing the things you need together makes these things less tightly coupled and allows shifting requirements easily while not impacting other's that use the composition.

OOP has its use cases and I think it is great for people getting started in development. It is also good for very small projects with limited uses. Composing things is the way to go for larger code bases.

[spark temple]

Hahaha. Okay.

[peak acorn]

Yea I never done real oop at work

My job entirely C 🤡

[twilit beacon]

can anyone tell me why the sleepfor function is getting squiggly lines? ```cpp
int main()
{
for (;;)
{
triangle();
std::this_thread::sleep_for(10);
deletor();
}

return 0;

}

i have the included threads but the std:: gets squigglylined

[azure mica]

"becomes necessary very quickly to manage increasing complexity that similar things need to be grouped into a single bucket so that you, the programmer, can easily manage what's going on"

OOP is the increased complexity.
Thankfully the industry seems to be moving away from OOP as a programming paradigm and back into functional.

OOP is great and all, but it has a lot of drawbacks that we simply shouldn't have to deal with.
This, in my opinion, becomes especially prevalent once things get big.

[cyan niche]

Like what?

[peak acorn]

Have u tried checking what it complains about

[azure mica]

https://github.com/readme/featured/functional-programming

[twilit beacon]

yeah, i looked on stackoverflow but noone got anything similar

"no instance of function template "std::this_thread::sleep_for" matches the argument list"

[silk eagle]

https://tenor.com/view/ah-shit-here-we-go-again-ah-shit-cj-gta-gta-san-andreas-gif-13933485

[cyan niche]

This article indicates that modern languages pick up functional principles into their own design. You could successfully argue JavaScript relies on it.

[silk eagle]

so you're saying

[azure mica]

JS is great, cuz it lets you write code.
OOP can be great but you need to bow down and lick its boots first.
Then you need to continue doing so for the next decade.

[silk eagle]

everything is functionally functional

[azure mica]

Also OOP usually has higher complexity meaning higher refactoring and maintenance cost.
I'd argue that the main success of OOP comes from the success of the languages that implement it cough cough Java.

I remember reading this a while back.
And I think some of the arguments presented here are really good.
https://suzdalnitski.medium.com/oop-will-make-you-suffer-846d072b4dce

[twilit beacon]

ok nvm fixed it

[azure mica]

Node Version Manager fixed it?

[twilit beacon]

no, i added cpp using namespace std::literals::chrono_literals;

[azure mica]

Then why are you talking about nvm ? 😂

[twilit beacon]

nevermind

nvm == nevermind

[azure mica]

Okay if you don't want to talk about it.

[twilit beacon]

str nvm = "nevermind"

[azure mica]

Yeah I don't think this is true

As a matter of fact I am pretty sure it evaluates to false.

[twilit beacon]

1

[azure mica]

jk, I am messing with you
😄

[twilit beacon]

reminds me of the time when stewie from family guy took anabolics and trolled brian

[nocturne galleon]

Im currently learning about arrays and objects, it doesn't seem too hard learning about it but I'm afraid of having to put it into use

[cyan niche]

What specifically about OOP makes you "lick it's boots first"?

[trail remnant]

This reminds me of this talk: https://www.youtube.com/watch?v=QyJZzq0v7Z4

The basic thesis is that "norm" basically equates to popularity, and then he goes into what made the top ten languages popular. The basic idea is that for some reason, functional programming languages never ended up hitting one of these reasons for very stupid historical reasons that basically came down to random chance and raw luck.

Then the presentation goes off the rails half way through, where things basically come down to very obscure hilarious oddities like Java becoming popular because they spent billions of dollars on marketing, and the fact that Objective C being "object oriented" happened because Smalltalk was on the cover of a popular magazine at the time, and that inspired him to follow in the footsteps and make everything an object.

The real killer point of the presentation is that functional programming is a "style", not a set of features. You can write programs in a functional style in nearly every popular language, including Java, the most anti-functional language around. Even better, since that presentation came out, Java has been aggressively adding functional features that allow some very functional styles to exist in Java. Support for streams and stream processing basically brings Lisp style coding to Java. Java also now has lambdas, which are the quintessential feature that allows you to skip writing custom classes to implement things like event listeners. I'm not a fan of Java, but it's hard to argue that Java is a "pure object oriented" now. It's a multiparadigm language, like basically every single popular language now days.

As a Clojure guy, he theorizes that some functional languages might end up in the top ten most popular languages, but ultimately he says at ~42:14

Classifying languages as "functional or not" is kind of arbitrary, definitely very fuzzy, and ultimately not as important as talking about the style and their support for the style.

[lone hornet]

this is what my prof. marked me down for on my quiz. Am I wrong?

[midnight wind]

depends on the convension

GB can be 1024 megabytes

[lone hornet]

yeah I don't think the prof. specified it. I know that one kibibyte is 1024 bytes but when I got my quiz back he said that 1 kilobyte is not 1000 bytes

[mortal sigil]

https://physics.nist.gov/cuu/Units/binary.html

as per international standards: a kilobyte is 1000 bytes and a kibibyte is 1024

people generally use "kilobyte" when they actually mean "kibibyte" though

grading someone down for it, however, is pretty pedantic imo

[azure mica]

Interesting. Will give it a proper watch later but a lot of what he says seems to be on point.

[lone hornet]

@mortal sigil Thanks for the link, I'll reference it when I'm emailing my prof for the points i'm due for

[mortal sigil]

in the real world, if someone says "A gigabyte", they mean 1024 megabytes 99.9% of the time

so it's worth keeping that in mind

it's still on your professor for an unclearly-worded question though

[silk eagle]

google is wrong which is probably where the confusion is

bing is also wrong

theyre all wrong, i dont know why this has been a thing for so long

[mortal sigil]

no, this is correct. Kilobyte != Kibibyte

see the above link

[silk eagle]

oh

i thought it was the other way around

[mortal sigil]

Kilo: base-10
Kibi: base-2

[silk eagle]

well shit I need to go fix some stuff

[mortal sigil]

But again, people nearly always mean "kibi" when they say "kilo"

[azure mica]

Scientifically speaking you are not wrong.

But because of how binary works the kilo, mega, giga prefixes were adapted to be in scale of 1024 instead of 1000. Because binary really doesn't like the number 1000, since we can only use 0 or 1.

It is considered a consensus.

[mortal sigil]

IMO, "kilobyte" should be base-2, and "kilobit" should be base-10

that's how they're used like 99% of the time anyways

[azure mica]

Maybe but we are talking about bytes no?

[mortal sigil]

it's just a really pedantic issue. even IEEE admits that the usages are VERY fluid. If you're being graded, or if it's mission-critical, the number base NEEDS TO BE SPECIFIED

[silk eagle]

on an unrelated note, I had a long list of items that I spent 5-10 minutes pasting into <option> tags because I wanted to do it the cool way with multiple cursors and all that and kept messing up

but it was stylish

[trail remnant]

That's one of those arbitrary divisions of units. Computers love powers of two, so most computer units move in powers of two. So usually, computer storage is measured in 2^10, which is 1024. Except when they don't, because it's much easier to market things in increments of 1000, and measuring the size of a storage device it in only 1000 instead of 1024 means you can put a slightly bigger number on the box.

[mortal sigil]

It's only 2.4% off anyway, so for marketing purposes it's "good enough"

[peak acorn]

Kilo kibi is such a pain in the ass

If you ever gotta be very clear, N x 10^3 or N * 2^10

[silk eagle]

or just say kibi and blame it on the other person if they get it wrong

[peak acorn]

Big tru

Also you can't use bit. Kibibit etc is important to keep separate

[mortal sigil]

it's one of those issues where it's not worth arguing at all

like if a manager or professor wants to use it a certain way, might as well abide by their weird system

No amount of SI-unit-pedantry will EVER convince me to use "nepers" though

[peak acorn]

Wtf is a neper

[mortal sigil]

it's a logarithmic unit, like the decibel-- it comes up a lot in any math-heavy engineering

[peak acorn]

Ah

I ain't no scientist

[silk eagle]

im of the opinion that the definition of "foot" should change depending on who our king/president is.

measure their feet, average the measurements, boom thats your foot measurement for the next 4 years.

[azure mica]

I agree to some extent.

[mortal sigil]

I guess the best description is:
KIBIbyte: clearly base 2
KILObyte: implied base-2, techincally base-10
kiloBIT: clearly base-10

[cyan niche]

First, the idea that functional programming didn't take off because of luck is... Remarkably grasping. The reason why C and later C++ was so dominant was not "luck", it was because it was the best tool for the job at the time it was created, and still is in its realm of application. The idea that something like Haskell or Lisp could currently replace C in the area of embedded system is laughable: it's execution speed and compiled size could never approach C on this front, despite its claims to the contrary, C is king in these regards. Lisp (Multilisp, interlisp at the time) also required a specialized virtual machine which in those days meant specialized hardware. Unix was built on C and so this was also a major factor in preferring C over other program languages. And the reason Unix was written in C was because it was originally written in Assembly and the port from Assembly to C is a lot less trivial since C is far, far closer to the actual machine code than functional programming.

This is not luck, and not Oracle's marketing, but simply that C was the best tool for the job at the time that was most practical and efficient. This also prevents functional programming's adoptions today. Functional programming terms are hard to grasp, OOP is easy and intuitive. The tool itself, whether it be Clojure or Haskell is difficult to use and the people that know how to use it are less prevalent therefore more expensive.

For the record, I like functional things included in my programming languages, and I suspect we'll see them blurred more as the best of each designs are included into programs. I certainly think some functional programming concepts could be incorporated into lower level language design and improve things fundamentally, and I don't really care which language you like the most. But the idea that "OOP was a trillion dollar mistake" narrative gets old on me. It gets rehashed but when I ask for specifics it's always, "Well OOP makes things difficult", but no one so far could show me how exactly in everyday programming it actually does. And don't say recursion because that's a limitation of imperative rather than OOP design, nor a problem that I face on the day to day.

[trail remnant]

Did you like, watch the presentation, or did you just make a quick emotional response to my bad 30 second summary of a complex presentation that covers decades of programming history?

For reference, Oracle didn't spend a billion dollars marketing Java. Sun microsystems did, which is part of the reason they went bankrupt.

[cyan niche]

I'm responding to you, not the video.

[trail remnant]

I was summarizing the presentation because I thought it was interesting. Your comment belongs much more on the youtube video I linked. I wasn't sharing my own thoughts, nor did I make any comprehensive statements about specific programming languages.

For reference, here's how I started off every single one of my comments:

The basic thesis is
Then the presentation goes off the rails
The real killer point of the presentation
As a Clojure guy, he theorizes

Also, here's a reference from 2003, about how Sun spent the money, not Oracle.
https://www.theregister.com/2003/06/09/sun_preps_500m_java_brand/

[cyan niche]

No I agree with them about Oracle.

or uh, Sun

And if I had to choose to program in the rest of my life in Haskell or Java I'd pick Haskell easy.

[trail remnant]

Go watch the presentation. It has a surprisingly good explanation (with plenty of references) about how the history of many languages was inspired by what was popular at the time. And many of the things that were popular were based on luck. He explains a surprisingly compelling line of reasoning about how Modula (a functional language) could have been the inspiration behind Objective C, if a butterfly had flapped its wings just a bit differently.

Do I agree with him? Mostly. He's got some more extreme views on functional languages, but I think the general idea is pretty compelling.

If nothing else, go listen to just the one question/answer at 42:14. He explains that it's really hard to define what a "functional language" really is. There's no generally agreed upon definition for what it means to have a functional language. If it's just "what features does the language have" then basically every modern language qualifies to be functional.

As far as the performance argument goes, Rust would like a word with you. An immutable by default language that provides strong compile time guarantees and matches C in raw performance, with the ability to directly target specific embedded hardware. There's no reason functional programming styles can't be blazingly fast.

As for the "easy to understand" that's arguable. I find functional concepts are easier to implement, design, and explain. Strong Object Oriented Java style XML documentation always gets difficult to read, and pushes people into pointless arguments over how to specify class hierarchies that usually end up making code less flexible and more confusing. I'd spend a great deal of time arguing that the only "good" Java code is the code that's operating on streams, which is basically functional programming with more steps and a less friendly language.

The really funny kicker is that C isn't object oriented at all. That's the whole reason C++ exists in the first place. (Which you'd know about the intermediate language between C and C++ if you watched the presentation)

I was never on the "trillion dollar mistake" rant. I don't think object oriented programming was a critical mistake. I'm from a time period after the billion dollar Java marketing hype train. The environment has changed a lot, and I don't see any extremely harmful ideas coming from Object Oriented programming, when done in moderation. (This applies to everything. Anything done in an extreme way is dangerous) The point of the presentation is that the lines are already blurred, and they always have been. Even some of the most ardent OOP guys will admit that composition is better than inheritance, and at that point it's just a pedantic argument on what programming style you like.

Languages are tools, and picking the right tool for the job always requires a careful and thoughtful analysis of the requirements you're working with. I do have strong personal preferences on which languages I like, but I'm still using GoLang at work regardless of how many hours I could spend ranting about critical design mistakes and annoying problems that really upset me.

[spark temple]

OOP isn’t a language. It’s a design pattern. So, comparing it to JavaScript isn’t something you should do.

[azure mica]

Ofc, it's just when I think OOP I think Java so I might have used it somewhat interchangeably.
My bad.

[spark temple]

Not really. Functions are just a way to structure code. You don’t really need more than one function for an application really. Just becomes unwieldy.

[spark temple]

No worries. Just wanted to make sure people understand appropriately.

[azure mica]

What's your take on OOP vs Functional?
@spark temple

[spark temple]

I mean, the main reason is with functional programming, it had a high performance cost on the CPU due to context switching. L2 cache wasn’t as big as it is now and the predictive execution wasn’t as good.

[spark temple]

Use the right tool for the right job. There’s balance between the two.

Yes, even in the same code base.

[azure mica]

Fair.

[spark temple]

Nothing is really inherently bad in terms of tools and patterns we come up with. It’s just how they are implemented is typically bad or not the right one for the problem at hand.

[trail remnant]

Definitely. At the time when C (and A/B that came before) was created in the late 60s, computers were dramatically slower. They couldn't afford any high level abstractions. I don't even know if they had L2 cache on those old mainframes. But C has been showing it's age for a long long time now. There have been over 30 years of attempted C replacements because C kinda sucks as a language. When comparing C with today's languages, C is basically glorified assembly code. You only really want to write C when you're dealing directly with the hardware, and you really really need to not have any abstractions between you and the hardware.

[azure mica]

Yeah like using Python for everything...

[spark temple]

All other languages are typically built on C. Nothing wrong with it. Again, pick the right tool for the problem at hand.

[trail remnant]

I'd argue that outside of rare situations, inheritance is something to be avoided. Similar to how goto is treated today. There's a good reason why experienced developers don't have code littered with complex class dependency chains.

[trail remnant]

Originally? Yes. But even Rust has been self hosting for a while now. Most highly performant popular languages are self hosting.

The only real cases where you want to use C/C++ is when you need to manually manipulate lots of memory, which is inherently dangerous. There's a reason that only one out of the top 10 most popular languages allows raw memory manipulation and doesn't have a garbage collector. For 99% of typical use cases, you really don't want to be manually managing memory.

[azure mica]

I mean these days Rust seems to fill the C/C++ gap pretty well.
Tho its community could focus more on actual programming.

[spark temple]

I did say favor composition. But the phrase is composition over inheritance. A closed loosely coupled system is typically good.

Though there are use cases for inheritance, few though. Better to define an interface and compose the implementations.

[trail remnant]

I think most of the power of Rust comes from its community, and their general focus on high quality over quickly churning out things to meet deadlines.

[midnight wind]

ik people who really only write in C

[spark temple]

It’s not really dangerous if you know how to count. It’s just that it’s hard for most people.

[midnight wind]

yeah, you have to know the computer a lot more

[azure mica]

Maybe... I have own thoughts about the Rust community.

[azure mica]

Yeah and us Programmers really don't like doing the hard stuff.

[trail remnant]

I think we should separate "bad" from "dangerous". "Bad" implies some moral wrong, or something that never has any merit whatsoever.

"Dangerous" meanwhile, explains the inherent risks associated with certain activities.

[azure mica]

There is a reason why we don't program in binary.

[trail remnant]

No programming languages are morally wrong. But there are definitely lots of things that are highly dangerous.

[spark temple]

That’s fair.

[azure mica]

Oh my... u just gave me a perfect idea.
Ima make my own most offensive and immoral programming language I can think of. Just to spite the Rust guys.

[midnight wind]

that's just skill

[trail remnant]

I'd argue this isn't about the "hard stuff", but more about being practical. If you make a single typo (something that is bound to happen eventually) then you're opening yourself up to basically everything going horribly wrong with binary. But with a compiled language, (even C) it at least protects you from most typos.

[azure mica]

Yeah and all it takes is to take this way of thinking one step further.

[trail remnant]

No matter how skilled you are, accidents happen. The aviation industry is one of the safest in the world because they admit that certain things increase the risk of danger, and should be avoided even if they aren't inherently wrong.

[trail remnant]

I'm not the one advocating for extremism. That's you.

[midnight wind]

well yeah, but that's what testing and debugging is for

[trail remnant]

Hol up.

[midnight wind]

also proper coding practices

[azure mica]

Neither am I. But if we are going to draw a line somewhere why not push it in either of the directions? What's a bit more.
And then u find yourself at the end of the scale.

[trail remnant]

Are you actually arguing that it's possible to prevent all bugs by just being skilled?

[midnight wind]

did I say that?

Bugs will happen regardless of language

[azure mica]

I think that's mathematically impossible.
Something along (not exactly) these lines...
https://en.wikipedia.org/wiki/Halting_problem

[peak acorn]

just be smarter

[trail remnant]

Let me ask you this then. What are you actually advocating for? I don't even know what your argument is. You've just been arguing against whatever I'm saying.

[midnight wind]

Just saying that there are valid reasons to use C

[trail remnant]

I never said there weren't. So why are you arguing that?

[peak acorn]

C awsom

[azure mica]

Assembler awsom

[midnight wind]

The only real cases where you want to use C/C++ is when you need to manually manipulate lots of memory

[trail remnant]

bnry awsm

[azure mica]

Voltage awsm

[trail remnant]

Relevant:

[azure mica]

Let's just go back to Analog Computing. Problem solved.

[trail remnant]

Butterfly programming is the best.

Did we seriously become a meme from 2013?

[azure mica]

I like the idea of analog neural nets.
https://youtu.be/GVsUOuSjvcg

I wonder how far we could take it.

Would it be better than digital computing? Who knows.

[peak acorn]

what do u even do programming an analog computer

whats the use case

[azure mica]

Neural nets.

[trail remnant]

I doubt we'll do a lot of that. Digital computing is far too useful and has far too many general uses to be replaced overnight. Just look at how slow it took huge sections of the industry to move from X86 to ARM, and that just required writing a new compiler and recompiling programs.

[peak acorn]

Then it'd just be a coprocessor
doesnt roly get programmed

[azure mica]

Also Analog Algorithms.
It's the same as Quantum computers. You need quantum-specific algorithms.

I don't really know much about this field. It's outside my scope of expertise but I think the idea is interesting.

[trail remnant]

Coprocessors are more likely, but even then it's unlikely that we'll ever make a hard switch away from digital processors as the main control unit for computers.

[azure mica]

Oh yeah for sure.

[trail remnant]

There are actually many dead ends in the history of computing. One of the most interesting is the alternative set of computing hardware that used cryonics instead of transistors. Source: https://spectrum.ieee.org/dudley-bucks-forgotten-cryotron-computer

[azure mica]

But then if we had a General Purpose AI, would it matter?

[trail remnant]

Literally the untimely death of one man in 1959 might have changed the entire history of our computer hardware.

[trail remnant]

General purpose AI is a pipe dream that's not going to happen in our lifetimes.

[azure mica]

I think we are currently limited by the hardware. If we had a big enough (assuming same speed at big scale) memory unit with sufficiently powerful (assuming speed again) computing unit I think we evolve a strong AI.

[trail remnant]

What's interesting is that Dudley Buck's cryo computers could have paved the way for a smooth transition to quantum computers if he hadn't died and had continued to make advancements. But after he died the microchip revolution pulled all the funding out and the advancements stopped. It's just a theory, but his death might have pushed quantum computing back by one or two decades.

[celest junco]

I’ve been pretty impressed by the recent progress in AI with GPT3 and Dall-E 2. Even the dumb GitHub copilot is pretty impressive (though less so).

[trail remnant]

I've had many discussions with AI believers, but I remain unconvinced. Most of their arguments seem to fall back to the idea that Strong AI is an inevitable thing that is guaranteed to happen eventually (the philosophical idea of naturalism). I find the evidence for such a claim to be lacking. Many many things need to go exactly right, and many many things need to not accidentally fail.

[midnight wind]

The rate at FOSS AIs developed vs Dall-E is impresive

Stable Diffusion

[azure mica]

This is nothing. Based on the AI industry and research papers we are about to make a lot of big things in the field of AI.

And Stable Diffusion is awesome.

[trail remnant]

The AI art generation is definitely impressive, but there's no evidence that Neural Networks will create a general AI. I'm not arguing against neural networks. Those will continue to be used in more and more places, and disrupt our lives a ton. But they aren't generalized AI. The way they're trained is designed to create very specific networks for specific things. There's zero evidence that an AI designed to generate text would be able to suddenly learn how to play chess without manually retraining the network, which makes it not a general purpose AI.

[azure mica]

You are correct, but the fields of NNs etc (weak AI) and strong AI research are very much disjointed.
Can't really compare one against the other.

Also we are really starting to move past the basic NN design. So I think we should just refer to weak AI as AI and not NN.

[trail remnant]

So you think there's upcoming research for Strong AI that will suddenly change the status quo? Most people I talk to can't even conceptualize the difference between strong and weak AI, and those that can still can't point to any evidence that strong AI are anywhere close to becoming reality.

[azure mica]

Oh no, in my message above about breakthroughs I meant weak AI.
As I said before it seems that we are very much hardware limited in terms of developing a strong AI.

[trail remnant]

No bike shedding. I don't care what we call it. Based on historical context, most people who talked about AI a long time ago were referring to strong AI. It was difficult for them to imagine a computer that could create art, but couldn't hold a conversation with someone, and couldn't understand human feelings.

So I'd argue that we should really always refer to NN as weak AI.

But that's not a useful topic. Pick a definition and I'll use it. I'm not here to argue about definitions of things that don't exist yet.

[azure mica]

Well the field of AI has evolved so much in the last decade... a lot of terminology has changed.
It also relates to what is known as an AI effect.
https://en.wikipedia.org/wiki/AI_effect

[trail remnant]

I agree. I can't remember where I found it, but there was a compelling argument that barring some insane breakthrough in quantum computing, we're at best around 100 years from being able to fully simulate a human brain at full speed. (which was the hypothesis that once we're able to do that, we can just throw a brain in a jar and have a strong AI)

[trail remnant]

Such is true in most fast moving fields. See the above debate about OOP, and how the original creator of the wording doesn't like the new modern definition.
Also, see "agile".

[azure mica]

Well my problem with calling all of weak AIs neural nets is that the NN design has evolved so much, calling it a NN might be an oversimplification.
It's like calling a bee a bug. You are technically not wrong.

[azure mica]

I am actually very much against the "replication" idea.
I think that strong AI should have the traits of human brain but should not simulate it. Otherwise it's not really a strong AI is it?

[trail remnant]

Yes, there are big advances in NN designs. But I don't think it's wrong to have a broad category of classification. There's a good reason that biology has multiple layers of classification for different things. In the tech industry, we don't really have the idea of broad categories for things. A language is either OOP or functional. We love to think about things in black and white thinking, and we have a hard time thinking about things at a higher level and a lower level at the same time.

[azure mica]

To this day I have no actual idea what "agile" really is. It just seems like a term that is thrown around a lot.

[trail remnant]

I like the logic of simulation of a brain. It's much easier to make modifications to an existing design than it is to create a brand new design from scratch.

Once you start trying to recreate intelligence from scratch, you suddenly run into a giant world of philosophy that no one has good answers for.

[trail remnant]

Oh man, now we're stepping into my world. I've gone around and through this subject a TON.

You've seen the original Agile Manifesto, right?

[azure mica]

Yeah but simulating a real brain would be crazy expensive (computationally).
It might be easier at that point to just make your own.

[trail remnant]

Here's the original website, as published in 2001: https://agilemanifesto.org

[azure mica]

But yeah we'd need much stronger quantum computers with quantum-NNs and other quantum algorithms.
To this day I barely understand how they deal with all the probability.
I should read up on this.

[trail remnant]

The most interesting part of the article I read was not the computation part (which I understood very well) but it was the biology part. It turns out that the original estimations in the 70s (which haven't been updated) were WILDLY wrong about how complex the brain is. There are dozens of things the brain does that are mission critical for complex thought, that weren't even theorized back in the 70s. We now know about some very complex things brains to to "think" that kinda destroy the idea that NN could ever reproduce the same things a brain does. Like the fact that each neuron is actually a small NN itself. Individual neurons can generate non-linear inputs, which wasn't understood at all in the 70s. They just knew there was electricity in there. There's also complex chemical reactions that change how neurons respond to different thing across wider regions of the brain. There's no really any analog in NN design as far as I know. (Although some things the brain does are similar to current deep NN designs like backpropagation, but what the brain does there wasn't understood in the 70s)

If brains worked like we thought they did in the 70s, then yeah, some time in 2040 we'd see a computer that could model a brain. But it turns out brains are WAY WAY WAY more complex than anyone thought in the 70s. My current theory is that we'll need to understand how the brain works before we'll be able to build a strong AI. I'd watch developments in neural biology over computational developments if you're really looking for strong AI.

Anyways, on the topic of Agile, what happened was very simple. In the 90s, big businesses started trying to design software projects, and kept failing a whole lot. So there was a conference where a bunch of top industry experts (Some names you've probably heard like Martin Fowler. Check the list of names on the website) got together to discuss how they'd found success, and figure out what worked and what didn't work. They eventually settled on four basic principles that they all found they used in their unique processes.

The principles were simple, and not perspective.

[azure mica]

Well I tend to disagree with this methodology. I think that while strong AI is definitely strongly influenced by our biology it would be an entirely different (conscious or not) being compared to any living brain.
The problem is we can't really find a good starting point for such an approach.

I remember reading about an idea a while back, about creating a complex system of multiple flexible NN that can change their structure, each sub-NN in theory representing a portion of functionality that can either strongly or weakly connect to other sub-NNs.
Then you somehow have to measure the performance of the system which is not easy at all and encourage it to move towards a specific direction. Tho I think they found a problem with this and the idea was abandoned I don't know if it was ever pursued to begin with but it sounded okay, tho it was never presented in detail.

[trail remnant]

My point isn't that we'll simulate a brain, but that we'll need to understand how brains work before we can reproduce them. Given how little we actually knew about brains 50 years ago, there's a very good chance that there's a very critical important part to "thinking" that's missing from our AIs. It seems very likely that a deep understanding of the brain will lead to being able to reproduce "thinking".

[azure mica]

Either way I have a train to catch tomorrow early morning. Cya.

[trail remnant]

The idea of sub networks is great, but it leads to practical problems. How do you train those sub networks? Once mistake with a subnetwork, and everything starts to fall apart. Brains train themselves in a very different way than NN do.

[trail remnant]

One last thing, is that the Agile Manifesto and "agile" changed over time. It turns out you can't sell an idea to people. You need to sell some sort of consulting. So many different "implementations" of agile appeared as people rushed to sell agile to billion dollar companies. Those are things like "Scrum", which prescribe a process with about dozen weekly meetings as a solution to "people over process". It gets about as dumb as you might imagine. Most people talking about Agile are running into the conflicts between the ideals of the manifesto and the realities of running a for-profit business. That's where most of the stupidity happens. If you really want to find people who know what Agile really represents, always ask about the manifesto. It's never been bad advice, even if it doesn't tell you how many mandatory weekly meetings your team needs.

[azure mica]

I heard of the manifesto, will make sure to read it.

[cyan niche]

No I hate the manifesto.

Everyone uses a different flavor of agile, but there's no reason to have some sort of creed for it. Any two companies aren't using the same processes.

I know because I used to do this professionally.

[trail remnant]

So it's your professional word against my professional word?

[trail remnant]

The manifesto is very basic. Do you really believe that people should put more effort into contract negotiations than collaborating with customers?

The manifesto doesn't say anything about mandatory weekly meetings or sprints or how to plan things or any of the traditional things people associate with "agile". I'd argue that the agile manifesto and agile are two distinct different things now.

[wind horizon]

Strongly disagree with this point on manifesto:

The most efficient and effective method of
conveying information to and within a development
team is face-to-face conversation.

Sounds like typical this is how I communicate best so others must be the same.

Me fighting with discords crappy markdown support. 😂

[trail remnant]

Remember that this was written before a time when real time video was possible. What other alternatives do you believe are the best way to communicate?

[trail remnant]

Discord markdown support kinda sucks

[wind horizon]

I don’t believe there is a single best way, people all communicate in different ways.

Personally I prefer written over verbal in 70%+ of cases. Similarly I think what is being communicated effects what medium works well.

I think it’s all it depends and imo that’s where agile should be today, continually improving on what works well for your team not trying to adhere to something specific.

[trail remnant]

My main point was that "agile" is now a product that's sold to businesses, with all the good and bad that comes with that. It started out as a thing many people with lots of industry experience agreed were good ideas, and slowly morphed into a product sold for cash with little regard to quality software and quality products. Scrum certification companies get their money regardless of if your company launches successful products.

[trail remnant]

So, responding to change over following a plan?

[wind horizon]

Yeah that’s true, I mean I have only worked 1 place that did certifications for it. Most other places I have worked were far more lax and didn’t buy scrum / agile stuff. (Besides say software)

[trail remnant]

Yeah, "Agile" is tainted and dead. The actual agile manifesto has been completely ignored and turned into a for profit product that's ruining our industry.

[wind horizon]

Well the entire purpose, imo, of agile is to stop letting process get in the way. What better way than taking your retros and saying, hey this isn’t working let’s try and improve it. Even if our improvement breaks out of the typical “agile” flow, but works better for the team.

[trail remnant]

The manifesto doesn't even mention retros. Those were added by a specific company that had a specific way of doing things that mapped really really well to business consulting.

[wind horizon]

I do think the typical agile we are used to today is a solid starting point for any team though. I just don’t think we should worry about does what we do fit into agile.

Yeah well it doesn’t have to be retro, I’m just using that as an example since it’s a common process used to review.

[trail remnant]

For reference, the actual manifesto says

The most efficient and effective method of conveying information to and within a development team is face-to-face conversation.

That's a bit different than "always use face to face communication". Maybe you aren't looking for the most efficient way to communicate information. Maybe text is fine because you're sending dry technical information and not emotional information about the quality of your product.

[wind horizon]

I think I’m more just annoyed with old co-workers I had that would complain all the time “X isn’t agile” or similar things. Always point to something they didn’t like and saying this isn’t correct agile. Imo they should stop worrying about that and instead voice why this is causing pain points and work towered solutions.

[trail remnant]

My whole point is that Scrum is a crappy excuse for the agile manifesto, and people are so tainted by scrum/kanban that they can't think of any other way to run a software team. Agile has become a fill in word for scrum.

[trail remnant]

Oh, definitely. That's the type of person who wants to have a strict set of rules to tell them how to think. That's what waterfall is for. Agile is the exact opposite of that kind of toxic stupidity.

Well, ideally. In reality, things are different. "Agile" means "scrum + what we like", and the manifesto is completely ignored.

BRB, getting food.

[wind horizon]

Yeah I just have heard it argued about so many times now that I’m grumpy about it. 😂

I have hit the point of idc wtf we call what we do, just does this process work and do we have a way to continually improve it. Lol

[deep flicker]

"Disregard 'Agile', acquire working code"

[wind horizon]

My current job has terrible agile processes, they improved like 10x over the past year though, so I’m still happy. There is a lot to still improve imo, but I at least know the system is working in the sense it keeps getting better.

[wind horizon]

Now you got me thinking what do I want for dinner. 🤔🤔

[trail remnant]

So you mean "working code over extensive documentation"?

Most of the stuff that we want to do as developers is part of the basic four principles. It's just frustrating how often people say "Agile" when what they really mean is "scrum" or worse, "our own in house process for creating the best code and you should take our process as law and not argue with it and also accept any changes from upper management without complaining".

[trail remnant]

Sandwich and soup is good.

[wind horizon]

Ordered some sushi haha

[trail remnant]

I think my point is that there's no such thing as "Agile process". That's the worst kind of oxymoron. The agile manifesto expressly says that you should value people plover processes. But the sales people selling consulting products have brainwashed people into believing that agile is a process, not a philosophy. Because you can't sell a philosophy.

[wind horizon]

But a sandwich does sound good

[trail remnant]

The whole "X isn't agile" argument is stupid. Arguing over what isn't agile is the wrong mindset. Agile is about caring about the right things and focusing on improving those and tracking the right stuff. Rather than argue about what isn't agile, they should be doing what a previous senior engineer did, and meet with the team to figure out how to reduce the number of meetings we have because they were disrupting development. He didn't sit around and complain about what was or wasn't. He did something about it to try and improve things. There was a discussion between team members and our manager, and eventually we figured out how to cut down on our meetings to give us more time to work on our product. Whining is very counter productive. I found that out at least three teams ago.

[trail remnant]

Picked up sandwiche and soup. But I got stuck in construction traffic. Making the drive how now.

[dapper flax]

The 12 agile principles do define what "agile processes" are. But yes, a core value is people over process. So a ceremony that provides a setting to get people talking has agile qualities.

[wind horizon]

Construction is the worst. I used Uber Eats, over charged but no construction haha

[trail remnant]

Do you actually want to argue that principles == process? I'm in the mood to go through every single one of the 12 principles and explain why none of them describe processes, if you're up for that.

[trail remnant]

My 20 minute food trip turned into an hour away from home

[dapper flax]

Maybe some other time. I just wanted to point out the fact that the priciples give two definitions for what agile processes are.

[trail remnant]

Strong disagree. Cambridge dictionary defines "process" as

a series of actions that you take in order to achieve a result.
Dictionary.com defines it as
a systematic series of actions directed to some end
In the context we're discussing, Merriam Websters defines process as
a series of actions or operations conducing to an end
(Thanks for being the most verbose and least practical dictionary websters.)

None of those definitions describe what's in agilemanifesto.org

The main page has four points, but says

That is, while there is value in the items on the right, we value the items on the left more.
Which is the opposite of "systematic". Nearly all of the 12 principles on the other page are statements, not prescriptive actions. The ones that are commands to do something have a deep level of flexibility, which makes them more "guidelines" than "process". For example,
Deliver working software frequently, from a couple of weeks to a couple of months, with a preference to the shorter timescale.
That doesn't tell you what you HAVE to do. Rather, what is beneficial to do.

[dapper flax]

Principle #2 defines agile processes as:

Agile processes harness change for the customer’s competitive advantage.

Principle #8:

Agile processes promote sustainable development. The sponsors, developers, and users should be able to maintain a constant pace indefinitely.

[trail remnant]

The word "process" does appear in the 12 principles, but the 12 principles are NOT processes. Do you understand the difference?

[dapper flax]

I think we are on the same page 🙂

[trail remnant]

Also, refer back to the main statement

Individuals and interactions over processes and tools
That is, while there is value in the items on the right, we value the items on the left more.
The point isn't to strictly define a process that "makes you agile". The point is that whatever your current processes are, they exist to serve your business, your employees, and your customers. And that those processes are always open to change. The whole "agile certification" thing is the exact opposite of that. It defines a bunch of processes that you team MUST follow if they're going to be agile.

Honestly though, I'm fed up with the corporate software development world. Like 90% of the teams there are operating like waterfall companies what pretend the mandatory meetings give them the secret magic sauce to have functional software teams. I'm just tired of it, so I'm looking at jobs working in a smaller team/department/company that's less prone to corporate mandates from on high.

[dapper flax]

Sounds like you might enjoy working for a startup

[trail remnant]

Definitely. I started off in a startup, but I was new, inexperienced, and had some pretty big disagreements with a really bad manager. I spent quite a few years in the corporate world, and I think I'm ready to go back with more experience.

[dapper flax]

Definitely important to have a good rapport with your manager. It's worth asking who the hiring manager is if its unclear during the interview process.

[celest junco]

Idk that start ups are exempt from agile dogma. I’ve worked at a startup where that was present and I work at a large f500 where no one mentions the word “agile”. We do sprints but they’re just a way of organizing work.

[wind horizon]

I have worked startups, mom and pop size, and all the way up to multi billion / F500 tech companies.

What I learned was the team and manger have the biggest impact, the most boring company can be fun and engaging with a great team.

I do sometimes miss the wild ride of the startup, the lack of rules / HR and everyone is just doing whatever. It’s like the Wild West anyone can go pioneer a new path in the company or rework huge sections of code / change frameworks just by putting in the effort.

Big business turns much slower, but also has the best work life balance imo. Phenomenal PTO, great pay, and other talented people you can learn from even if your later in your career are some great perks of big tech.

At the end of the day I’d always pick a great team and manager over big or small company.

My personal requirement as an IC is the manager needs to have at some point, preferably in past 5 years, been an IC themselves. I’m okay with further up the chain ppl being far away from IC or never being an IC, but direct engineering manager I’ll probably never take another role where they were never a dev in the past.

[trail remnant]

Absolutely. At the time I was new and kinda desperate. I don't regret taking that job, but I wouldn't do it again.

[trail remnant]

Unfortunately, right now I'm contracting, so I have literally zero PTO, and the pay is only "good". Right now our team is pushing to finish things on tight deadlines, so I don't really have lots of opportunities to work directly with experienced engineers. It feels like I'm kinda distant from everyone else on the team. Most of the reasons to be part of a larger company just aren't there at my current job. Fortunately, I'm very isolated from political BS, but I'm just looking for something different.

And I agree that it's pretty terrible when a manager doesn't have practical experience doing the work. It's too easy for people to forget what it's like if they've been out for a long time.

[wind horizon]

Oh that sucks, yeah imo that sounds like a situation I’d be looking around at new opportunities as well.

[lament bridge]

Is there somewhere a good video to learn how to create multi-filters in React, maybe in Javascript itself?

Mostly the filter is made of user input and select elements and dates.

[cyan niche]

I think it's an unrealistic statement. Businesses of sufficient size will begin to demand contract negotiation between its developers and stakeholders. A sprint itself is a contract negotiation in a manner of speaking.

In government spaces contracts are everything and they have to be, what tickets you're allowed to work on, how long the sprints are, what work is delivered, it's all negotiated within a contract. It has to be a contract because the government won't work with you without one.

I see where they're coming from but when you build a product and the client comes back and says, "No this isn't what we want" a contract is your way of proving that you delivered what was promised. A contract can be as simple or complicated as you wish.

Agile is the best SDLC imo because it's a framework that can be applied to almost any business demand. Almost every sector has embraced it as the de facto way of doing things.

[silk eagle]

I like to make the same project once a year just to see if I've improved any, and when the day comes that someone needs this specific project that makes Roku direct publisher feeds, I will have the best one.

ill probably just load it up with ads and post it to a mailing list

[twilit beacon]

just do it on a laptop/pc

[trail remnant]

So, you think government agencies that prioritize contract negotiations produce better software? 🤔

[hollow basalt]

Then have one

[cyan niche]

Government agencies require contracts as per legal requirement.

[trail remnant]

That wasn't the question.

Government agencies can totally prioritize contract negotiations is they don't care about the quality and cost/speed of their software. Inflated budgets and cost overruns are pretty normal for government agencies, so it's not exactly the sort of thing that makes your stock go up. (See Intel the last few years)

It's also worth noting that even when you have an obligation to negotiate a contract, the contract can be negotiated to hire an agile team to build as much functionality as they can in a given time period, rather than set requirements in stone before the technical experts have gotten involved. The Social Security Administration is already doing this.

Source from the US government: https://www.oversight.gov/report/SSA/Agile-Software-Development-Social-Security-Administration

[cyan niche]

I could give you a laundry list of problems wrong with governmental red tape, but suffice to say not much anyone can do anything about it because that's what's required from the government itself because there are so many regulations they have to comply to.

For example, there's reporting requirements for every single change that happens to the code base.

[trail remnant]

Specifically, the Social Security Administration is actively aiming to change their delivery method from one giant blob to a more incremental delivery method: https://www.oversight.gov/node/224354

[cyan niche]

And I use agile at a government agency.

We use sprints just like anyone else.

[peak acorn]

do you use agile or just use sprints to track work

[trail remnant]

You're confusing the ideal with the compromises different organizations have to make due to their situations. And you still haven't answered my question on if you think organizations that spend more time on contract negotiations produce better software.

[cyan niche]

I'm not about to get into a discussion gatekeeping agile.

[peak acorn]

am i getting time confused or have yall been arguing above functional programming and agile for 2 days now?

[cyan niche]

I got a life man

I gotta argue only in my spare time xD

[silk eagle]

is that ironic? i feel like thats ironic

[peak acorn]

no i stg at least some ppl have been arguing for days

[cyan niche]

¯_(ツ)_/¯

[trail remnant]

I posted a summary of a video, then I described the history of the definition of the word "Agile", and some people think that means I'm a fanatic because I used two words associated with misery and suffering for many engineers. That's what happened.

I've been asking clear questions to the people replying to me, and I haven't been getting any answers. The people who have been arguing with me can't answer a simple direct question about what they believe. I've stated many times what I believe, and they can't seem to directly disagree with me. 🤷

[silk eagle]

well

that has to be ironic

[cyan niche]

I don't have enough information to answer your question because there isn't any hard data to back up what "customer collaboration" vs "contract negotiation" actually is. I don't know what's better because I don't know what they are, let alone do I have any sets of results that adequately compare the two.

Second, the question is ill formed because it asks an answer to something that is irrelevant. Governments have to use contracts.

And if I were to put on my philosopher hat, any discussion between individuals when discussing a course of action on something is essentially, a contract negotiation, and each party at the end of it is expected to uphold to the agreed decision.

This could be as simple as deciding how to sort a list of product items by default to what kind of database to use. Any time a consensus is concluded an informal contract is formed, whether it be verbal or written.

What would customer collaboration even look like? Me sitting down in a meeting with some users about the software asking them specifics on how a button works or what color it should be? Whatever they decide is still essentially a contract between the two parties present. It seems to me what it actually meant by "customer collaboration over contract negotiation" is "informal over formal agreement", where the latter requires written and signed agreement to a list of stipulations while the former is a simple thumbs up to a slack message.

Which just goes to show you just how unclear the entire manifesto is, and this is why I don't like it.

The principles are solid, however, and provide clarity in their meaning.

[hollow basalt]

Ah yes hello world

[twilit beacon]

yo, can too much math in one line crash c++ programs? it always stops working immediately ```cpp
posX=rsin(the)cos(phi);
posY=(rsin(the)cos(phi))cos(A)-(rcos(the));
posZ=(rsin(the)cos(phi))sin(A)-(rcos(the));
int point = (scWidth(floor(posY)-1)+posX);
zBuffer[point] = '';

[twilit beacon]

how do i fix a segmentation fault

fixed

[twilit beacon]

i did this dot thingy

and now its this

funny circle doing whateva

[hollow basalt]

As we can see above. This is what slowly getting crazy looks like when learning c/c++

[twilit beacon]

i did it, it is finally smooth

[trail remnant]

Priorities. It's about priorities. Did you read the one paragraph of text?

That is, while there is value in the items on the right, we value the items on the left more.

They don't say "never negotiate contracts". They say you should put more effort into collaboration than trying to win contract negotiations.

If you haven't experienced this before, there's an entire industry based around trying to manipulate people. Here's a random primer I found: https://www.bookchums.com/books-on-manipulation/

This type of thinking was far more popular in the 90s when the manifesto was written, and the manifesto stands in contrast with much of the conventional wisdom of how to run a business by being a manipulative jerk who takes advantage of everyone they can.

The important part is collaboration over a toxic negotiation where both parties are trying to cheat the other side out of as much as they can. Those businesses and people still exist, and that type of thought process doesn't lead to quality software.

[astral ether]

Posting that seems manipulative

[trail remnant]

It is manipulative and evil, and it's the sort of thing some people do. That's the contrast between people who don't collaborate and those that do. That's the context that's important to understand when you try to understand the difference between the software development methods of the 90s and the agile manifesto.

[deep flicker]

Sorry, my comment was being snarky 😁 . I've dealt with places where people who've never written a line of code talk about how they are 'scrum masters' and 'agile experts' and plaster on layers of bureaucracy when at the end of the day someone needs to write something that makes the 1's and 0's going into the CPU correct.

TBH most of the stuff I work on now is small enough it can be managed via a simple "epic > story > tasks" categorization

[twilit elk]

https://www.youtube.com/watch?v=bB340S0tGf8

[deep flicker]

Hey if the process isn't working, we can just pay for more training to become a Level 2 Certified Scrum Master ®️

insert quote about doing the same thing over and over again here

😄

[sly marten]

I don't know C++, but I would say; yes and no - the math itself probably isn't the problem, but how you've done it could be
afaik the computer will still break up the math into pieces anyway, it's more about formatting/readability in the IDE/code files

someone please do correct me if I'm wrong, as I'm pretty much just guessing off of my understanding of how computers work

[hollow basalt]

Computers are literally made to Compute. Yes it breaks up the operations into something that the architecture can do.
So not too much math, but wrong math would crash the program

[twilit beacon]

i already followed up with a message that fixed it

[hollow basalt]

it doesn't have to be

[sly marten]

fix it, yes, but fixing isn't always a direct answer to a question 🙂

[twilit beacon]

i just asked if the math could have been the problem here for the crash, then i downloaded gdb and saw it was a segmentation fault and applied fix

[hollow basalt]

nozemi is right though

[twilit beacon]

yes but the question wasnt whether computer languages are designed to be able to do math, whereas it was more meant as "is this possibly the reason my computer couldnt handle and it therefore ended with a crash"

[sly marten]

yeah, that's fair, but since I only know the "how do a computer work" part of that question, I could only give you tips that will help regardless of what you do and which language you program in 🙂

so in knowing that, you also know math operations in one line isn't a problem, it's something else that causes a crash - which you already figured out though

[twilit beacon]

i probably should have expressed the question a bit clearer

[sly marten]

if you wanted a more specific answer, probably 😛 but regardless, I always try to give answers that are more useful in the long term - sort of like building a stable foundation before you build the house on top

because by knowing fundamental things, it's also a lot easier to troubleshoot other things 🙂 though that way of answering people's question means a little more work and thinking on their part (which is my intention tbf)

[hollow basalt]

yup tbf

[keen hound]

The only thing that can cause an exception in math is division by 0

[midnight wind]

arcsin.. arctan..

Out of bounds

[keen hound]

That depends on the OS, on windows floating point exceptions are disabled, so it won't crash your program

You're just gonna get NaN as a result

[midnight wind]

I guess yeah

Just remembering what gives me errors on calculators

[waxen river]

Our school district has blocked GitHub because “It could be used nefariously”

Any ideas?

[midnight wind]

about what?

why? or getting it resolved?

if you want to get it resolved, talk to admin, and say how you use it for programming

[waxen river]

Tried by both me and our AP computer science teacher

Didn’t work…

[midnight wind]

what did they say

worst case, use a vpn or another git server hosting service

[peak acorn]

Yeah unless u wanna cause a real stink ur gonna be sol

[waxen river]

Npm isn’t working right because of this

[midnight wind]

I would write how it disturbes learning and you can't develop

[hollow basalt]

Vpn time

[spark temple]

I’d used a vpn in the meantime but for sure reach out to the network admins. Yes, GitHub can be used to distribute bad code but… professionally I can’t work unless GitHub is working. I’ve left work in the past when it has gone done because I couldn’t do my job and there was no point in staying.

[wind horizon]

Also maybe a simple proxy will work, unless they are doing packet inspection likely can just proxy your traffic. Lol

[waxen river]

Breaks code of conduct

[wind horizon]

You asked for ideas on getting around the schools blocking, but are worried about following their code of conduct. 🤔😂

[hollow basalt]

Laws should be freeing not the other way around

start the revolution

https://tenor.com/view/riot-south-park-chaos-chainsaw-protest-gif-20878334

[misty thorn]

Hell yea overthrowtheplanet.org

subtracts ".org"

What if all websites didn't have extensions

No .com no www.

Just... getshitdonedotcom

[hollow basalt]

but you just said no .com

[misty thorn]

It's all spelled out hehe

[midnight wind]

it's possible

http://ai./

[misty thorn]

Let's go I'm ready for meta web

[spark temple]

If this was for a job, then totally understandable. Since you pay them, I don’t see the ethical dilemma. They’re preventing you from using one of the most effective tools you’ll use in everyday software development. Seriously, there’s very few better tools than GitHub.

Not just using for your code but for documentation of modules you’ll need for almost every single programming language.

This isn’t an ethics thing where you are taking money from a job and need to play within their security playground because that’s what they think keeps them safe. This is harmful to your education not having access to GitHub.

[bold scarab]

I have some questions about js frameworks and web dev if anyone can help?

[celest junco]

They’re probably just blocking DNS from resolving if I had to guess. Adding entries in /etc/hosts would fix that

[hollow basalt]

no one

[polar arrow]

shoot 😄

[hollow basalt]

curry for 3

[bold scarab]

I am trying to learn svelte and I dont understand when I am supposed to use html or the .svelte file

[polar arrow]

always use .svelte

[midnight wind]

.svelte, since it's not valid html

it needs to be compiled first

[polar arrow]

yeah, use svelte/kit

[bold scarab]

I am using both

[polar arrow]

wont work

[inner wraith]

VPN to home or a cloud provider/TOR, probably the former because websites block TOR annoyingly

Some arguments aren't worth having

Or use cellular

4G hotspot might be a viable workaround

[waxen river]

I am on a school computer

[deep flicker]

Boot a Linux live USB + VPN?

[inner wraith]

A desktop?

[waxen river]

Until I can get a replacement laptop battery its not an option, currently I have like a 40 min battery life

[waxen river]

Shity mac with an RX 580

[inner wraith]

Failing that just copy your repos+deps to a flash drive and take it home to do pulls/push

[deep flicker]

Could also init a repo on a USB drive then plug into another computer to sync with github https://stackoverflow.com/questions/20628420/how-to-make-usb-pendrive-as-a-git-repository

[waxen river]

gitlab also still works

[inner wraith]

Pretty much what I just proposed

[waxen river]

for now, so I can fall back on that for syncing, hard part is packages

it fucked up NPM to error every time due to the proxy

[inner wraith]

Repo on flash drive, npm at home

Or fix your stuff

[waxen river]

OH WAIT, I am locked out of my gitlab account at school

[hollow basalt]

you don't have to be

[waxen river]

take a guess why

[inner wraith]

Sounds like your solution is to fix your problems lol

[waxen river]

If you guessed I made my GitLab account using my github account You would be CORRECT

https://tenor.com/view/banging-head-bang-gif-12070477

Wait I can change it to a username and password

Crisis Averted

[urban flint]

crossposting a cool linux development thing I did here as well 🙂

#linux message

[trim warren]

can you go back to previous scanner?(like going back to main menu) using a string scanner?

[urban flint]

Is that addressed to my post? You can switch layers with key codes yeah

[trim warren]

oh no it's a problem I've been working on

[sturdy ingot]

You should at least add the programming language (and possibly even the framework) to your question.

[hollow basalt]

Pepperoni

[twilit beacon]

working on this cool project, will add hotfixes tomorrow but for now its good ```cpp
#include <iostream>
#include <math.h>
#include <windows.h>

float cons[] = {108, 36, 0.05, 6.28, 10, 0};
char bu[3888];
const int mix = (cons[0] / 2);
int miy = (cons[1] / 2);
int posX;
int posY;
int posZ;

int main ()
{
while (1)
{
memset (bu, ' ', 3888);
for (cons[5];cons[5] < cons[3]; cons[5] += cons[2])
{
for (cons[5]; cons[5] < cons[3]; cons[5] += cons[2])
{
posX = cons[4] * (mix / miy -(cons[0] / pow ((mix / miy), 5) * 2)) * sin (cons[5]) *cos (cons[5]) + mix;
posY = (cons[4] * sin (cons[5]) * sin (cons[5]) + miy);
int point = floor (cons[0] * (posY - 1) + posX);
bu[point] = '#';
}
}
std::cout << "\x1b[d";
for (int x = 0; x <= 3888; x++)
{
putchar (x % 36 ? bu[x] : '\n');
}
}

return 0;
}```

[peak acorn]

What does it do

[silk eagle]

i personally prefer to write paragraphs as variable names

[bitter edge]

is this only for coding

[silk eagle]

its for most forms of software/hardware development-related things

[bitter edge]

so I can ask questions about a PCB

[peak acorn]

Prob won't get as good a reception as other topics, but yea u can ask anything lol

[bitter edge]

alr so I wanna try to create a little pcb that has a micro usb port and a led and I'm wondering how much resistance I would need for it

[deep flicker]

This is what breadboards were invented for! https://en.wikipedia.org/wiki/Breadboard

You should absolutely prototype the circuit to confirm your layout, component choices, and functionality before making a PCB.
I would recommend getting a breakout board for the header (e.g. https://www.adafruit.com/product/1833 for uUSB), a couple of LEDs (never hurts to have spares), and an assortment of resistors at different values for testing.

The most important thing to consider when designing a circuit is power dissipation. The datasheet posted says the diode can dissipate 150 mW, with a forward current of 100 mA. For the following I'm assuming a series circuit with only the diode and resistor as the load, i.e. +5 V |---Diode---Resistor---| GND

Using Ohm's law, V=IR, (https://en.wikipedia.org/wiki/Ohm's_law) and the power law for circuits, P=IV, (https://www.allaboutcircuits.com/textbook/direct-current/chpt-2/calculating-electric-power/) we determine how much voltage can drop across the diode as

P = I * V --> V = P/ I --> 0.15 W / 0.1 A = 1.5 V

assuming the USB header give us +5V, the voltage drop across the resistor is then

5 - 1.5 = 3.5 V

and since it's a series circuit, the same current flows through the resistor as through the diode (100 mA), so it's R value is

R = V/I --> 3.5 V / 0.1 A = 35 Ohms

[deep flicker]

The values for the LED current you have are quite high though, usually 15-35 mA is more reasonable https://www.adafruit.com/product/754, https://www.adafruit.com/product/301

[deep flicker]

You should also check the current rating on whatever you're plugging in to the USB header to make sure it can deliver enough power https://resources.pcb.cadence.com/blog/2020-what-are-the-maximum-power-output-and-data-transfer-rates-for-the-usb-standards

[bitter edge]

i never knew such a little thing would take so much time to plan

[trail remnant]

Nah, I totally get this. It's frustrating to me as well. I worked on several highly dysfunctional teams that didn't know how to communicate or build working software. They were scrum teams with certified scrum masters, but they couldn't deliver software on any sort of a reliable schedule. It was kinda pathetic, and deeply frustrating for me as someone who enjoys seeing people use my software.

It's why I'm pretty pedantic about the difference between today's "agile" and the "agile manifesto". The two are different things, but the manifesto is often treated like it's just as toxic as some of the agile garbage that often happens to teams. The manifesto is just a document, and it's written by industry veterans, and surprisingly reasonable. I don't think it should be followed like a religious textbook, but the advice often borders on common sense. (As a software team, your main goal should be to make software. Much advice. Wow.) Meanwhile, today's "agile" is just a corporate product. It's a product sold for profit, and it's usually much easier to find a crappy ripoff product than a high quality product. I don't think that agile has to be bad, but much of it is pretty terrible. However, it's important to have perspective. Decades ago, waterfall projects could literally spend years negotiating contracts and detailing out every single minor detail, only to start the project and realize that the most basic assumptions were wrong, and the entire planning phase had been wasted. Yes, I've had multiple support incidents this year because another team pushed unexpected changes that broke our team's code, but I'd take that over spending a year planning and not writing code.

[bitter edge]

public class Main {
    public static void main(String[] args) {
        System.out.println("Hello world!");
    }
}
``` i em coder

[deep flicker]

Yeah, proper agile manifesto style development is a better methodology since it's self-correcting/easier to detect when an assumption is missing or wrong. Iterative approaches to design/development are much less likely to waste massive amounts of time

[deep flicker]

Do you have soldering experience with surface-mount parts? If not I'd strongly recommend a through-hole design as a first PCB instead of surface-mount. It's waaaaaaaaay easier

[bitter edge]

through-hole design?

[deep flicker]

https://en.wikipedia.org/wiki/Through-hole_technology
https://en.wikipedia.org/wiki/Surface-mount_technology

[bitter edge]

that's about as much as I know (it failed)

also pretty sure I burnt it

[deep flicker]

The resistor in your screenshot is a surface-mount resistor, which are tiny and not easy to solder by hand. Same with the pins on the USB header

[deep flicker]

Yeah those header pins are bridged

[bitter edge]

I'm getting a heat gun with some solder paste to fix that issue

also as far as I know my little usb thing is responsive

it just doesn't work...

[deep flicker]

Yeah because the power and ground pins on the USB header are bridged, so it's shorting out the USB port

[trail remnant]

Unfortunately, you sometimes get "agile" teams like the ones that I was on, where they had a retro, only for people to complain about the same things every single retro, with zero changes ever made to actually fix those practices. (Because heaven forbid we slow down for one or two weeks to make our team move faster the rest of the year.)

[bitter edge]

Also zed are you familiar with ribbon cables

[deep flicker]

From https://www.electrokit.com/uploads/productimage/41016/41016968-1.jpg, this is what you want no solder between the pins

[bitter edge]

That's why I'm getting a heat gun and solder paste..

it will make that much easier

[deep flicker]

Not the super tiny ones like you find in laptops, I usually just use these https://www.mikroe.com/ribbon-cable-40-wire-male-male-20-cm

[bitter edge]

oof

[deep flicker]

Or just solder a JTAG header and break out the pins on the other side

[deep flicker]

You can get them(https://www.digikey.ca/en/products/filter/flat-flex-ribbon-jumpers-cables/457?s=N4IgTCBcDaIE4EsBGSD2A7ABAYwIZIBsBTEAXQBoQA2KUBAEyhAFoBGABgkoAcAXJkJV4BPbiUgh6AZ2wgAvnKA), but they're super finnicky and the board connectors are almost impossible to get correctly attached by hand

[deep flicker]

The complaining will continue until morale improves 😛

[wind horizon]

I feel like this is legit 90% of retros. I have seen very few things improve from them, ideally I think retro needs to be formalized a bit after “complaining” is done and some action items can be taken.

But it’s almost always just a spot devs can vent and make them feel herd then brushed off. Lol

[trail remnant]

True. I think if the expectation is that improvements only are allowed in a bi-weekly meeting, there's something wrong with the culture. It's also worth noting that 90% of teams are complete crap. (sturgeon's law and all)

My current team is pretty good, but most of the important improvements have come out of non-retro discussions. There's a general willingness to speak up and offer improvements to our processes, which is really great. The most impactful improvements usually happen when someone runs into a big problem in the middle of the sprint, and they come up with an idea to fix that problem, and bring it up during the next standup. Generally, if you communicate well, it only takes ~10 minutes to explain a problem, discuss your potential solution, and commit to a plan of action, so we can have that discussion after our ~8 minute standup. Our team is very supportive, so everyone is there to contribute improvements to ideas, not to tear them down.

Dealing with problems as they arise lets people feel appreciated and relive pressure throughout the sprint, which means those feelings aren't bottled up, so the retro isn't just a complaining session. We also structure our retros to explicitly have positive sections, which helps because our team team doesn't have systemic issues dragging us down, so most of our pains are temporary. The way our retros are run makes them sometimes useful for being a place to bring up time independent ideas with all the team members present and no distractions. Generally though, the main process at the retro is to create a story for future action. The retro shouldn't be your only tool for improving your team.

Unfortunately, most teams have major systemic issues and are constantly in pain, so the retro is just a regular meeting to uncover how screwed up you team is and how little things are improving, which tanks team morale.

[next cipher]

yeah I'm in this weird position where my regular job is a non-production team, but one of our big ongoing research areas has kinda suddenly evolved into an actual product and they want it out there now

so I'm helping pitch in on a fairly standard agile/scrum ish team at the same time as we're putting together/spinning off that team to actually become a product team

it's a very weird situation and i still have my regular job to keep track of and not lose track of shit (even though the total workload is pretty flexible)

[broken pendant]

hi.. im new to programming... is there an application that allow me to create simple GUI for python?

[bitter edge]

i better stop plugging this in ;- ; this thing almost broke my external hdd

[trail remnant]

Hardware design is hard. Now you know.

[bitter edge]

yep

part of me is telling me this will work and part of me is telling me this will fail

[trail remnant]

Honestly, I think the biggest problem with most teams is impatience. The business people push for things as fast as possible, without understanding the tradeoffs that will happen. There are other systemic problems, certainly, but not taking time to do things right is definitely one of them most common and easily avoidable ones.

[trail remnant]

Don't you still have to solder the plug to the board anyways? If you have zero solder skills, then it doesn't matter what the design of the board is like.

[bitter edge]

i have some solder skills

[bitter edge]

some of my finest handy work 🙂

[trail remnant]

[bitter edge]

yeah yeah yeah but do you know how small that is in reality

[trail remnant]

https://tenor.com/view/im-not-convinced-not-convinced-unconvinced-skeptical-doubt-gif-11697472

[bitter edge]

and I'm getting a heat gun with some solder paste to fix that issue

[trail remnant]

I'm not saying it's easy. I have literally only touched a soldering iron once in my life. But if it's out of your skill range, then it's out of your skill range.

[bitter edge]

I mean maybe if I had a fine point tip

but mine is like 1/8th of the board itself

[trail remnant]

Watching Louis Rossman rants definitely qualifies me to let you know that you need more flux. More flux fixes everything.

[bitter edge]

I don't have anything to really magnify what I see right now

[bitter edge]

what do you mean by "plug"

[trail remnant]

The board is green, and you're trying to use the holes to make an easier connection, right? The problem is you still need to solder the red pins to use the breakout board.

Instead of needing to solder this stuff yourself, you can always just buy a breakout board for the components you need. It's pre-soldered, so you just need to use the large pins to connect things.
Like this: https://www.adafruit.com/product/4090

[bitter edge]

I mean I can solder the red pins

just not with a soldering iron

[trail remnant]

You gonna solder it with your smoldering good looks or something?

[bitter edge]

no

i am gonna use a heatgun with some solder paste

[trail remnant]

Do you have a way to apply the paste?

[bitter edge]

it's paste

[next cipher]

yeah we recently got told by the product/business people "yeah we decided this is going to a small trial next month" meanwhile we are just barely getting to a demo-able state

like tech demo, not even remotely close to a complete product

[silk eagle]

speaking of demo-able states, I spent the day rewriting things line-by-line (it's python so its really only a few hundred lines), fixed dozens of bugs accidentally, and improved the speed of my service enough to have to artificially slow it down so it doesnt look like something broke.

I also had to make a save button that changes its text to "Saved!" for a few seconds and doesnt do anything because people didn't understand that things were automatically saving.

[bitter edge]

"it's python so its really only a few hundred lines"

[silk eagle]

well yeah each line in python is like 35,000 lines of C

[bitter edge]

i do java

[silk eagle]

what does that indicate? i have like 0 knowledge of java other than java webstart being gone so idk

[twilit beacon]

it prints out a circle, but this obfuacated code doesent work rn.

[bitter edge]

I think you're maybe exaggerating a lil bit

[twilit beacon]

java is even worse

[bitter edge]

how so

[twilit beacon]

class HelloWorld { public static void main(String[] args) { System.out.println("Hello, World!"); } }

[bitter edge]

and

[silk eagle]

well I mean a few hundred lines is still not much to rewrite in python

[silk eagle]

now make a web request

[bitter edge]

i may be stupid or smth but isn't that kinda javascripts job

[silk eagle]

its a lot of things' jobs, javascript is just the most direct way of doing it (sort of)

[twilit beacon]

its just stupidly a lot just telling the computer to print out a hello world. at least its just printf in c

[bitter edge]

but can you make minecraft plugins in python

[silk eagle]

you can make minecraft in python and then make python plugins for the python minecraft

[bitter edge]

Who's gonna do that

how big of a message are you typing

[silk eagle]

😎

[bitter edge]

usually people who code can atleast type pretty fast

[silk eagle]

I type like 130-150wpm whenever I'm sprinting

[bitter edge]

you can't sprint while typing

[silk eagle]

but if theres a bracket I fall down to 90 due to the whole python thing

[bitter edge]

java has more brackets

[silk eagle]

python has no *curly braces

[bitter edge]

they use this [ or what

[silk eagle]

its indent-based

[bitter edge]

press tab then boom you have an indent

[wind horizon]

Yeah I think retros often become the complain about things outside our teams direct control. Product making crap tickets, designs missing, etc etc. Things that should be action items to improve with those steak holders, but yeah. Lol

[trail remnant]

Give the demo, and then after the demo, tell them about your plan to give a perfect copy of your laptop to all the customers that want to use the software.

[bitter edge]

is the laptop an added bonus for free 🙂

[silk eagle]

test the demo in a VM, sell the vm image

[bitter edge]

the board is blue

[trail remnant]

This is actually a classic problem. Turbo Tax has a similar thing, where people don't understand that calculating taxes is actually trivially fast for a computer. So Turbo Tax literally sits there for seconds, showing you all the things it's calculating with little pretty loading bars, just to make sure everyone knows they did the work. While it may seem counter intuitive to intentionally make your software slow, think of it more as an accessibility feature that helps people understand what your software is doing.

[trail remnant]

Did you not see my MS Paint scribbles?

[bitter edge]

i saw but the board is blue your drawings are irrelevant

[trail remnant]

Tell the company that to deploy the software, they'll need to hire a full time person to sit there and set up new laptops with your exact same setup so everything works correctly. I'm sure management will love this plan, and rush ahead with deploying software that way.

[bitter edge]

how do you like my new design

even smaller

[trail remnant]

It's not blue anymore

[bitter edge]

yes now it's white

time to learn how to create 3d models ;- ;

[bitter edge]

just one more question : should I make this even smaller

[bitter edge]

alr your silence has answered me

make it larger

done

[bitter edge]

even larger

friends

[trail remnant]

Darklife just over here writing fanfiction

[sturdy ingot]

Yeah, I'd recommend you to write a non-obfuscated version first that actually works and then obfuscate the code. Bonus points if you keep it C only.

[twilit beacon]

i do have non obfuscated code but i for some reason cant multiply 2 values from 2 different indexes in the array etc. so ill have to figure that one out, besides im just learning c++ so i wont do c

[sturdy ingot]

k. But that's not C++. It's C with std::cout.

[twilit beacon]

but its a cpp file

how do i then make it c++

[sturdy ingot]

You'd write proper and clean C++ code; which would kinda defeat the whole obfuscation journey, right?

[twilit beacon]

i am making a clean version, but when that one will be finished, ill obfuscate it

this was just a test to get a rough idea of the obfuscation process

[sturdy ingot]

Yeah, that's fine. Still it's easier to start with a working variant and then obfuscate it, especially if you're learning a language. I'd also recommend you to use version control (e.g. git) in case you haven't tried that yet.

[twilit beacon]

what does a version control do

[sturdy ingot]

It allows you to have a look at earlier variants of your software/code without saving first_try.c, second_try.c, final_try.c, real_final_try.c, real_final_try_1.0.c, real_final_try_1.0_final.c.

By the way, if you want to have some ideas for obfuscating your code, here's a slighty obfuscated variant of your goal (C, not C++): ||main(c,d,e,f){for(f=0xf,c=-f-1;c<f;puts(""))for(d=-1-f,++c;d<f;)printf("%c",c*c<=f*f-++d*d?'X':' ');}||

[twilit beacon]

man whatever that obfuscated part is, it looks like pure mess💀

[bitter edge]

welp it's only 6.10$ to buy 30 of them

imma do it

[sturdy ingot]

That's the intention 😄

[sturdy ingot]

But if you're asking what that code does, well... it creates a (filled) circle.

[rancid nimbus]

Does anyone know how to resolve this?
error: 'strstr' is not a member of 'std'

[sly marten]

what language is this?

my guess is Python or C(++), but really no idea...

[rancid nimbus]

Cpp I think.

[twilit beacon]

how the hell did you decipher that bruh

what did you do to make it that short

there is no cos, no sin no nothing

[sturdy ingot]

Use lots of stuff that are "somewhat" valid in C code 😄

[sturdy ingot]

If you have a circle of radius r, any point (x,y) is within the circle if the Euclidean length of (x,y) is less than or equal to r. That's the same as sqrt(x*x + y*y) <= r. Square both sides and you have x*x + y*y <= r*r. Rename the variables and you get c*c + d*d <= f*f. Move the loops increment within the expression (which is, by the way, undefined behavior, but works fine on GCC), and you'll get c*c + ++d*d <= f*f. And then subtract ++d*d from both sides to make it slightly more confusing, and you'll end up with c*c<=f*f-++d*d.

[twilit beacon]

so if i wanted to make a "sphere", i could just get away with changing or fondling with x²+y²+z²=r?

pls not spoil everything if that is the case, ill try it myself first

just can you tell me if i got the gist of it

[twilit beacon]

oh i understand now

you used pythagorean theorem to get the distance

[bitter edge]

this is why i use java

like tf is all this math

[twilit beacon]

in java if you wanted to do the same you would use the same math

you would just express it differently

but my question now is, does it iterate like that without waiting until the for is done or what

[peak acorn]

That's a nasty for loop

[twilit beacon]

and what does the puts do

[peak acorn]

Print I think

[twilit beacon]

yeah but what is its signifficance

why it do what it do

[spring cradle]

newline?

[silk eagle]

well how else are u going to puts?

[twilit beacon]

but that would be \n or is this some speciality of c

[spring cradle]

i think cpp has puts

[peak acorn]

I don't understand that for loop at all

[twilit beacon]

i think it iterates trhough all the possible x and y

[peak acorn]

The condition is (x=-r-1)? So it continues until that difference is 0?

[twilit beacon]

to the put into the last printf statement where it calculates something

[spring cradle]

x or space if it's inside circle

[twilit beacon]

no thats first statement, the condition is x is smaller than r

[peak acorn]

Oh there's a comma

[twilit beacon]

wait ill post it again so we wont have to scroll up

[spring cradle]

does it take an r so the fn won't have to initialize a variable or something

[peak acorn]

Yeah they're stacked without braces

[twilit beacon]

wdym

[peak acorn]

1 sec

[bitter edge]

no I would find a library that simplifies it

[twilit beacon]

its just doing first for loop and in that loop there is the second one which holds the printf?

[peak acorn]

Yeah they are nested

[spring cradle]

yeah it's like a double nested one liner

[twilit beacon]

so c can nest without braces?

[peak acorn]

U don't need curly braces for for loops with one line

[twilit beacon]

oh its a one liner

[peak acorn]

This doesn't rly feel like c, what's with the parameter list?

[twilit beacon]

it is c just obfuscated

[peak acorn]

Our the function declaration entirely

[twilit beacon]

thats kind of the point of c coding

[spring cradle]

ooh i have one

[peak acorn]

didnt know that was valid syntax at all

[twilit beacon]

idk you dont need to declare main ig

[spring cradle]

#linux message

line 8

it uses a bracketed if inside an unbracketed while

[sturdy ingot]

Aye, exactly

[sturdy ingot]

Nah, that's obfuscation. Proper C99 code would look like this:

#include <stdio.h>

int main(void) {
    const int radius = 15;

    for(int y = -radius; y <= radius; ++y) {
        for(int x = -radius; x <= radius; ++x) {
            if(x * x + y * y <= radius * radius) {
                putchar('X');
            } else {
                putchar(' ');
            }
        }
        putchar('\n');
    }
    return 0;
}

[sturdy ingot]

Well... that's a Code golfing tactic, but more or less: you can skip the return type of a function and the parameter types; this will have C default to int. You will get a ton of warnings, but hey, you're not going for clean code with obfuscation there.

[peak acorn]

ah

any C code i write has hundreds of characters of warning config

[sturdy ingot]

Ah, I usually have -Wall -Wextra -pedantic active, at least in real projects. With C++, I usually have clang-tidy in some pipeline.

[silk eagle]

obfuscated C is just perl

for example, im pretty sure there's 1 variable here and its $line and the rest is just normal to perl

$line =~ /^\s*(\w+)\s+(\d+)/;

[sturdy ingot]

Eh, many other languages use the same syntax for regular expressions. Some of them don't have a special operator though and instead use a function call.

[sly marten]

lol... assuming one is made... it also depends what you want to learn and/or do... always relying on libraries certainly isn't always a good idea

also the performance of JVM is terrible in comparison

[sly marten]

back at this are we working on a Kotlin library for Minecraft server plugins

[peak acorn]

poggin

[sly marten]

class SetHomeCommand(
    override val name: String = "sethome",
    override val description: String = "Set a home at your current location."
) : HomesCommandBase() {

    override fun execute(player: Player, command: BCommand, label: String, arguments: List<String>) =
        player.createHome(if (arguments.isEmpty()) "default" else arguments.first())
}```
not too bad to define commands like this

but would still be nice if I could make kts work

[peak acorn]

kotlin weird

[misty thorn]

rate my samurai game fight theme (really early)

[bitter edge]

7.4/10

[bitter edge]

i just use java

what ide do you use ?

[sly marten]

oh, IDE, my bad, IntelliJ

[sly marten]

meh, until you get used to it, it is 😛 but it has a shit ton of convenient syntax sugar

[peak acorn]

never actually used it

[sly marten]

here is another proof of concept plugin

I had a whole translations plugin, so in my plugins I would do translation identifier keys to lookup translations off of - this let me translate commands, arguments, messages, etc etc (me being the server owner, not just the developer)

I also had a player settings plugin, so all of my plugins could use that plugin to store settings etc for players, which makes it easier to have settings for your plugins, but also easy to use settings across plugins

the kotlin plugin library will be open source once it's more complete, along with some of these plugins that will serve as proof of concept plugins

[slate frigate]

Feeling this a LOT as I'm building a genetic training algorithm

[bitter edge]

I need a anyone to give me 3 names for buttons

[bitter edge]

my when I work with css

anyone here use dreamweaver

if so how tf do I change this view

I don't wanna be staring at this like that. I want it somewhat vertical

[hollow basalt]

dreamweaver

[bitter edge]

what about dreamwaver

[midnight wind]

everything

[bitter edge]

what about everything

[midnight wind]

it kinda sucks

[hollow basalt]

[silk eagle]

I'm no web developer, but VS Code + some relevant extensions is probably way better for web development

when I do web development stuff, I just use this extension in vs code:

[wind horizon]

Honestly I didn’t know DreamWeaver still existed. 😳

[bitter edge]

i can tell

but geforce now just gave me a free trial to adobe creative and I wanted to try it

[glass phoenix]

@silk eagle live server doesnt work for php in my pc , what would i do?

[wind horizon]

I bet PHP has some dev server setup, usually all major frameworks have them. But if not or if you just want to run closer to prod you could setup a Docker file you run with compose that hosts it behind ngnix/Apache. Set the volume mount that way local changes will be replicated into the container.

[sly marten]

Button 1, 2, and 3

why would you have anchor tags around buttons though?

[sly marten]

2010 called, they wanted their editor back!

in all seriousness - I have no idea what DreamWeaver is like these days, but I doubt a lot of people use that... Try VSCode instead

[sly marten]

you need something that can host php scripts, the browser won't understand php

Live Server (afaik anyway) is for front end development, and not having to reload the browser every time you make changes to the code

[twilit beacon]

what would you advise me, start doing c to then obfuscate or continue with c++

[silk eagle]

to make buttons href

[sly marten]

but in fairness, it's probably because they're new to HTML, and didn't learn it's not the way to go

[silk eagle]

yes true, they are probably new to HTML. now I know the right way, but for the people here who dont, what is the right way?

[sly marten]

<a> = anchor tag = a link taking you somewhere (traditionally anyway)
<button> = a button = triggers an action that does whatever (usually executes some JavaScript)

[silk eagle]

oh so just use the <a> stylized to look like a button, understandable

[sly marten]

https://www.w3schools.com/tags/default.asp
https://developer.mozilla.org/en-US/docs/Web/HTML

these are two great resources for web development 🙂 documents the standard that browser engines follow (at least try to)

[silk eagle]

theres this one vs code extension somewhere that I uninstalled accidentally, it does some fancy stuff where whenever u write css it autofills browser support stuffs'

[sly marten]

you can certainly do that, I've done that many times - but things like Google, embed generators (such as whatever Facebook, Twitter, Discord, etc uses), accessibility tools (disabled people - mostly blind I guess) might treat buttons and anchors differently

anchors can also take you somewhere else on the same page (a button can too, but would need JavaScript to achieve that)

[sly marten]

yeah, not sure the browser support stuff is very necessary these days, it was much more so when IE11 and older was still used

though Webkit (which is what Safari uses) has some kinks, some specific to iOS (maybe iPadOS too), some specific to Safari 😛

[sturdy ingot]

That depends on what you want to achieve. Do you want to learn proper C or C++? Then pick one and learn (with proper learning resources, especially for C++; most online tutorials suck and only teach "C with classes"). Write lots of programs and get them reviewed, for example on https://codereview.stackexchange.com.

Do you want to write obfuscated code, but not go the way of Cthulhu with C++ templates? Then pick C.

[twilit beacon]

I just think that getting experience with programming before college (while also ofc focusing a lot on school) is just what i need to hopefully have a good future. This is really mostly to master a computer language and understand computers in depth. I was also thinking of maybe learning assembly, as that requires deeper, in depth knowlege of these systems.

[sturdy ingot]

Hm. If you want to have more in-depth knowledge as in "closer to the real system", then I'd recommend C, especially if you want to go into Assembly.

Also, its fun to see C code and the resulting Assembly next to each other: https://godbolt.org/z/9fxYMK1Mj (but that also holds for other languages like C++, Rust, Zig...)

[twilit beacon]

then ill do c, thanks for the help

if i hypothetically wanted to also use an ascii code (such as 0x20 for 32 for instance) instead of numbers, would that also assign it a value like C did in the obfuscated code

[bitter edge]

bruh I use that as my main editor lmao

[bitter edge]

^^^

[wind horizon]

This message ❤️. Please don’t use buttons as links. It’s one of the most common mistakes I see when ppl work in JS frameworks like React where they toss on click handler to redirect. Now your user can’t open the button as new tab, copy link, accessibly is shit, etc.

It feels like modern JS frameworks have made it too easy for ppl to never learn proper markup and you start to see things like div styled like button with an on click handler to open a new webpage. 😂

[sly marten]

just build the entire site with div tags

[silk eagle]

ok now hang on

I feel attacked

[young elbow]

when you think about it, everything in html is a glorified div

[latent valve]

fr

[wind horizon]

Nah there is a lot of differences how browsers handle specific elements, especially when it comes to accessibility.

[sly marten]

glorified divs, but the browser (like glamdrink was on about) treats them differently in terms of accessibility, according to HTML spec (which is a defined standard)

[sly marten]

not necessarily something wrong with this

[bitter edge]

why not use spans 🙂

[sly marten]

why use spans?

[bitter edge]

idk I've just used spans

and they atleast seem to work exactly like divs

[peak acorn]

By default divs are blocks and spans are inline iirc

[wind horizon]

I’d suggest you skim through an HTML course / guide or YT videos. Basics of HTML aren’t too big / complex so you can probably blast through it on the weekend and you’ll find it makes things a lot easier when you have some of the html basics down.

[bitter edge]

who are you talking to

[hollow basalt]

glamdring is right though

[wind horizon]

Well technically anyone, but was specifically referring to your comment on spans working like divs. 🙂

[bitter edge]

spans

[bitter edge]

also they do

[wind horizon]

Span are inline elements, while divs are blocks.

[bitter edge]

they still work the same though

[twilit beacon]

isnt nav better for navbar or still div or whateva

[sick fossil]

idk what category this question would fit but quick question:

Does mere existence of trojans on the system (Have not opened the files yet) cause issues? I'm trying to learn some reverse engineering and its the first time I try downloading malicious files. On a VM of course

[peak acorn]

The binary executable sitting on your system wouldn't be able to do anything unless something runs it

[hollow basalt]

that's a computer question

[twilit beacon]

computer programming

[peak acorn]

Computer

[silk eagle]

https://www.youtube.com/watch?v=jeg_TJvkSjg

[twilit beacon]

puter man

[peak acorn]

Segfault

[lapis briar]

can someone with knowledge of c help me? Can i use a switch statement with every case in a seperate sourcefile?

[peak acorn]

What is the problem exactly?

[lapis briar]

I dont know how to use multiple source files

I want to have one file for every case

[peak acorn]

Oh I think I understand

Wouldn't it make more sense for each case to be a function? Either way you need to compile all the files

Maybe search up simple makefile or just gcc them all

[sly marten]

yes and no, you can use methods in switch statements, where each method is defined in another file

you don't need knowledge of c to be able to answer this question tbf

if c has other ways to do it, idk, but I doubt it

[lapis briar]

thanks ill look into it

im very new to programming 😅

[sly marten]

np, but please try to avoid gatekeeping people from answering your questions, because I have 0 experience with C, yet I can answer your question just fine

same with questions like "anyone EXTREMELY experienced in PHP?" to then wait for someone to respond, and the question is just "how do I hash and unhash passwords?"
this is just an exaggerated example fwiw, but the point should be clear 😛

ask your question, without putting any arbitrary limitations on who can answer 🙂

So in this case:
"Can i use a switch statement with every case in a seperate sourcefile?" ->
"In C, can I use a switch statement where each case is handled in separate files"

[lapis briar]

ahh thanks, i get it

Shouldve thought of that

[sly marten]

yes, but also no 😛 you can't know until you know, you simply just asked your question and was focused on that rather than how to communicate it as good as possible 🙂

[lapis briar]

thats right, good to know for the next time though!

[sly marten]

hence why I explained it to you ^^ and also everyone else who comes across it

[nocturne galleon]

so, i'm working on a makefile, i want my main target to depend on a phony target that creates an output directory (for all my objects and such)
if i just do

discord: setup $(OBJECTS)
    $(CC) -o build/discord-shmem.exe $<

it complains about setup not being a file, which i suppose makes sense, but there is another issue, even if i filter-out setup (which please tell me if there is a better way to do it), there are no objects in $<, however if i flip their places ($(OBJECTS) setup), it works fine but then it doesn't create the build directory early enough

what am i doing wrong?

OH wait i am dumb

i should be using $^

[silk eagle]

Python 3.11 is out, finally we have fancy error messages that point things out (...at runtime)

and Exception notes

and whatever this is which you can use for whatever you use it for

https://docs.python.org/3.11/whatsnew/3.11.html

Most notably: https://docs.python.org/3.11/whatsnew/3.11.html#pep-659-specializing-adaptive-interpreter

[sleek blade]

I made a fancy play/pause button, thought I’d share

[bitter edge]

where is the download to this fancy play button

[sleek blade]

haha, download to what exactly?

Just a little project I'm working on, nothing released publicly, yet

[bitter edge]

looks good

[sleek blade]

Thank you

Came out a lot better than I was expecting 😂

[hollow basalt]

Why

ARR GEEE BEEE

[candid hare]

hey champs, im tryna figure out how i could possible make certain audio outputs change on a shortcut.

right now i do it by opening sound mixer options and changing it individual but would be nice to be able to do it off a desktop shortcut

anyone have any ideas or anything that could help?

[sleek blade]

Lol, why not 😂

[sly marten]

per application?

https://learn.microsoft.com/en-us/windows/win32/desktop-app-technologies my guess would be to look at this

[nocturne galleon]

Can anyone point me in the direction of a community for helping to learn javascript? I'm currently doing a bootcamp, and there's a slack channel for it but it would be nice if there was another resource I could alternatively use for help

mostly looking for people who I can talk to about it rather than hard learning resources, I have plenty of those

[marble forge]

try the Odin Project

[wind horizon]

I think some of the ppl that sell learning resources also have slack or discord servers, for example I think Wes Boss had one. No idea how populated they are or quality of ppl isn’t hem but prob worth a shot.

[echo notch]

TIL: In python 3.8+, they removed the ability to unpack a dictionary item in a lambda function:

ltt = {"linus": 0, "luke": 1, ...}
dict(filter(lambda (k,v): 'linus' in k, ltt.items()))

you have to handle it like a single tuple object:

ltt = {"linus": 0, "luke": 1, ...}
dict(filter(lambda k_v: 'linus' in k_v[0], ltt.items()))

[deep flicker]

https://www.youtube.com/c/EngineerMan/videos

^ This guy is a programming/development polymath

[vestal spire]

is there a place where i can deploy my create-react-app and add password protection for free? Like vercel's password protection, but I need the pro plan for that on vercel

[nocturne galleon]

You can host anything for free on oracle cloud, though as a vps it requires a tiny bit more management from you compared to managed services

[next cipher]

unlikely to exist without a catch. marginal cost of hosting a static react app is essentially free but password access requires actual server side verification, a database, etc etc and suddenly costs money

oracle cloud is truly free as a loss leader but the tradeoff is that you have to maintain & secure it yourself (also that you have to deal with oracle)

[wind horizon]

I think you can have free auth from okta back in the day for something like 10k accounts, not sure if that’s still the case. I think Firebase used to have free tiers as well, again not sure what changes have happened over the years in pricing.

If you can work with a 3rd party like that todo what you want you can get mostly if not fully free tier, but it’s going to require some mix and matching probably from different providers free tiers. 😂

[sly marten]

doesn't need database, the web server has a built in authenticate stuff

well, some web servers have anyway, but not sure of any free hosts - aside from configuring your own VPS though

[midnight wind]

Otka sso?

Because ik azure AD is free for a certain about if objects

But that still requires a backend

[sly marten]

they probably don't want to mess about with SSO or Azure either - at least the impression I get is they want something rather trivial

fairly sure the easiest way to get it for free is one of these free Oracle VPSs and configure it on there, then install Nginx and use that for password protection

[midnight wind]

Yeah

[upbeat aspen]

hey i can't seem to get streamlink running can someone help

[next cipher]

where are you storing your usernames and passwords if not in some form of database? doesn't specifically need to be a mysql server or whatever, but it has to go somewhere

unless you just want a hardcoded basic auth password sitting on the filesystem but that's hardly considered authentication

[spark temple]

Uh, I mean... that's what it ultimately gets saved in but you can use OAuth providers. But yeah, saving user information but passwords? Not needed if they authenticate with something like Google, Facebook, Amazon, Apple, etc.

There's multiple ways to solve a user authentication.

[peak acorn]

why wikipedia have a newline here

[hollow basalt]

cause it is more than
half

[silk eagle]

i did that

[wind horizon]

They used to have a stand alone option too, which used their same sdk but could even do basic user/pass not SSO. I haven’t touched any of there stuff in few years so not sure what there is these days.

[midnight wind]

huh

[next cipher]

you can use public client flows in SPAs to authorize access to other APIs or resources but that can never actually block or grant access to the site itself

the point is that you either need something dynamic which costs the provider orders of magnitude more than static file hosting, or you're stuck with some kind of hard coded basic auth which isn't meaningful security

[peak acorn]

Why is python so bad

🤨

How can I make an enum, or similar concept that has some values associated with it. Seems like the enum package is borderline useless or required a bunch of boilerplate to make it useful

Ok apparently i just fucked smth up, but nonetheless i shouldn't need to google as much as I am to set up some basic enum stuff

[storm berry]

class MyEnum(enum.Enum):
  V1 = enum.auto()
  V2 = enum.auto()
a = MyEnum.V1

[peak acorn]

Yeah I ended up figuring it out

Annoying though, mapping those enums to a string, or attaching a function to them? Should be easier.

For the string, I just ended up using a dict, which feels real dirty but it doesn't matter

[storm berry]

there's MyEnum.V1.name which gives 'V1'

[peak acorn]

Right but not exactly the enum name

[storm berry]

or MyEnum['V1'] which gives MyEnum.V1

[peak acorn]

assumign you're supposed to name enums BIG_CASE or similar
can't display anything like that

[storm berry]

how would you do it in another language

[peak acorn]

in java enum is just a class with some constant guys attached, i like that
other langs are similar

Also good, in C, enum is just a renamed int. So there's no expectation of being able to do fancy stuff, and it's also dead simple

[storm berry]

like what do you want from this

Right but not exactly the enum name

[peak acorn]

Just re-styling something, say States.MAIN_MENU should be transformed into 'Main Menu' or whatever, for each enum

I just made a dict and mapped them

[storm berry]

you could do

class States(enum.Enum):
  MAIN_MENU = 'Main Menu'

then States.MAIN_MENU.value gives you 'Main Menu'

[peak acorn]

Right so an enum can only have a single value attached i guess
Or looking at how it's structured now maybe you can just attach a class

[storm berry]

yeah you could probably have a whole class as the right side

[peak acorn]

Point is I think it's not as obvious as it could be. Which I guess is sorta clear since class States(enum.Enum) is one of the goofiest declarations of any lang

[storm berry]

I think it works like a normal class

[peak acorn]

Right I think it's just a normal class and each enum is a static field?

[storm berry]

but by inheriting from enum it gets some qol stuff

[peak acorn]

by the looks at least, i dont know python very well, I almost exclusively use python as a terminal calculator

[storm berry]

maybe it's hacky, maybe it's cleverly keeping things simple, maybe it's both

[peak acorn]

If it wasn't required to be python for this class I would definitely not do it

It's a networking class, so we made little TCP client and servers. I guarantee you I'd be done faster if I did this in C, lol

[storm berry]

python is pretty good at tcp though

[peak acorn]

Probably cos I do C for a job, and not python, but nonetheless

I think they just pick python becuase its easier to debug and they really don't want to put student onto new languages

[storm berry]

I'd pick python over C for most networking tasks

[peak acorn]

Yeah probably for the best

But personally I could have finished this assignment in C a lot faster

Especially this goofy state management stuff I don't know very pythonic ways to handle it

[sly marten]

It goes in the config of the web server, so not exactly ideal, but it works, and is pretty much exactly what he needs

[sly marten]

Ah yeah, this is what I was on about. Indeed, it’s not authentication, but they were also just asking to password protect their application on a free host.

Presumably they just want to show the thing to certain people, and not be accessible by just anyone

If they actually want authentication, oh man, they’re gonna have to do a lot of learning I’m afraid 😛 There are a crazy amount of ways to do that. They probably also want to get a cheap host rather than free host at this point.

Hetzner is pretty affordable. DigitalOcean isn’t too bad either, depending on what you get. OVH is reliable, but they have a terrible user control panel, and as far as I know not particularly good customer support - though I’ve been with OVH for years, never had to use the control panel or their customer support (I’m also above average experienced, so I don’t need their support for my own misconfiguration etc though)

So that’s a little bit of food for thought, I can elaborate if I’m onto something they actually want.

[sly marten]

Fairly sure I’m right in my assumption above - they don’t even want you to see their react app before authentication has happened, without a backend, a web server with basic auth is best you’re gonna get.

Because otherwise you will have to check if they’ve authenticated for every request they make, before serving the React app

[sly marten]

So if you haven’t already figured it out for yourself, please read my messages above - and ask if you need 🙂

[nocturne galleon]

Tbh i would never pick python over C, it's borderline unreadable to me

[midnight wind]

It depends on where

But for scripts and stuff python 100%

Webapp, python

[nocturne galleon]

I would javascript or typescript or something, hell even bash would be better

Python is by far the hardest language for me to read

[next cipher]

yes i understand that - the point is that free static file hosts pretty much universally don't let you configure the web server or add anything (even basic auth) beyond just shoving the files down the pipe in response to any request, unfortunately

[next cipher]

but yeah this is what i said - a public client auth flow doesn't help it has to be server side

[silk eagle]

I feel like you've seen some bad or complicated python code then

[midnight wind]

yeah python for the most part isn't terrible

[nocturne galleon]

I've tried to read a fair bunch of projects

[hollow basalt]

oh no js/ts electron and/or their bloated projects

[spark temple]

I mean, if you want to lazy load you’re app is one way to make sure the package of the SPA is sent (aka your react app). Typically this is how it is done. A request comes in and then the token is verified as valid.

But the only other solution to ensure someone doesn’t get the SPA before authenticating is VPN and only allowing it on the network access.

Lastly, I think you’re talking about something else in regards to what I am. Cause I was referring to how users are verified aka; generating a token for login. You’re talking about authentication, aka does this user have rights to get this resource.

[sly marten]

I mean, if you want to lazy load you’re app is one way to make sure the package of the SPA is sent (aka your react app). Typically this is how it is done. A request comes in and then the token is verified as valid.
not sure what you're talking about here - sounds like something on some backend verifying an access token before it sends the files (in this case the React app), typically something such as JWTs - which I doubt is even close to what this person wanted, so it's probably just gonna confuse them more to start suggesting these types of things to them

But the only other solution to ensure someone doesn’t get the SPA before authenticating is VPN and only allowing it on the network access.
no - basic auth of a web server does this... before it serves you any files, it has you enter whatever credentials you configured in there

Lastly, I think you’re talking about something else in regards to what I am. Cause I was referring to how users are verified aka; generating a token for login. You’re talking about authentication, aka does this user have rights to get this resource.
well, I'm talking about what the guy who asked a question needed - or rather, what I assume he needed - a simple non-advanced way to show his React app to people he choose, not whoever has the link.. basic auth is effective for this purpose..

if he was asking about authentication and having a login/registration system on the other hand, that would be something else

[spark temple]

Clearly you don’t understand what I’m saying. Good luck solving the problem.

[sly marten]

lmao ight man... enlighten me then?

how convenient, blaming it on me lacking understanding, but refusing to explain? 😂 very mature, goes to show you're very confident 🤡

[spark temple]

Did I attack you? No, clearly we are not on the same page of what we are talking about.

I got what you said, I didn't say anything in regards to you being technically wrong. I was trying to address what I originally replied to.

But sure, I'm the immature one if that helps you sleep better at night.

Have a good one because I got work to do.

[sly marten]

my bad then - I interpreted it that way, but I will eat my words 🙂

thought we were all still trying to solve the original question lol

[hollow basalt]

Lol

[shadow hawk]

https://www.youtube.com/watch?v=xbaHVu2wzcw

[drifting narwhal]

Cool

[kind magnet]

i noticed they might be using python in their # development, its too easy syntax you need c like syntax

no thats a joke but honestly base c syntax is horrible

i never felt the need to cry when coding before

i am working now without any auto completion apps other than intellisense (which does not work for some reason)

[silk eagle]

is there a significant difference between

dict.__setitem__(key, value)

and

if dict.get(key):
    dict[key] = value

other than separating the logic behind checking the existence of the key? (so if you modify/use the same key multiple times you would only need to check its existence once, saving python precious time)

[spring cradle]

readability is a plus

[next cipher]

the difference is one is vastly more readable & idiomatic

underscores denote an internal method you should probably not be using directly unless you know exactly what you're doing and there is no other way

[wind horizon]

So if your worried about performance of a hashmap lookup you shouldn’t be writing your program in Python to start. Lol

With that said if you know you want to set this key anyway why not just use something like dict[key] = value instead of reaching into private method? 🤔

[storm berry]

if you want to check if the key exists in the dictionary it would make more sense to write if key in dict:

[silk eagle]

just cuz python isnt fast doesnt mean I shouldn't look to make it a little bit faster with reasonable compromises. (not saying it would be wise to use the private method, but point is there are a few approaches and I'm trying to learn best practice by comparing them)

[storm berry]

https://effectiviology.com/premature-optimization/

[silk eagle]

Fair point, but keyword in my message, reasonable. Like with if dict.get(key) and if key in dict, it's basically the same thing from a development perspective, but one could be noticeably faster and I'd never know if I didnt ask questions

and I'm not asking this to retroactively optimize existing code, just so I can be aware of it in the future

[storm berry]

also these two lines aren't the same actually

[silk eagle]

I'm trying to decipher that