pedagogy

[green warren]

And these people will tend to dive into code without really analyzing the problem first. It could be a personal thing, like they feel they need to prove something, or it could be from poor management at their previous job.

[subtle owl]

Oh there is no problem doing non-coding tasks.
We view software engineering as more than just writing code. I think here it's more due to being their very first job and being too used to finding all the answers on a platter from stack overflow. So they need more help to nudge them to think through things more

[green warren]

Real problem-solving skills are hard to come by and kind of hard to teach, the only real idea I have is to assign progressively larger problems to tackle, but you probably already know that.

What do they do when they get stuck and the answer isn't on Stack Overflow?

[subtle owl]

they just ask the other engineers, which in turns antagonize them (or they are too nice) because the question is very simple since it comes down to just looking where is a function called.

[misty dirge]

@vernal garnet your question was off-topic for this channel. See #❓｜how-to-get-help

[vernal garnet]

Thank you

[slender gale]

umm what is a subset

[alpine bear]

set inside another set. Can be a proper subset or just subset.

[noble aurora]

a subset isn't a set inside of another set, at least not the normal definition. if all the elements of a set A are in set B, then A is a subset of B. if they're not equal, then A is a proper subset of B

[weak mortar]

so guys after i finish learning basics of python
i just start learning Libraries?

[misty dirge]

yes and no, but mostly no. you should learn how to do different things (build a website or a Discord bot), and learn how to use the libraries as they relate to what you're trying to do.

[weak mortar]

True statement

[valid trail]

find small projects, then find larger projects

and if you choose a good variety of projects, you'll be forced to learn the libraries of Python

if you do enough projects, you'll learn those libraries pretty well

[weak mortar]

what's the most used ones

[misty dirge]

it depends on what you want to do. I'm most familiar with data science, and pretty much everyone in the data science space uses numpy, pandas, and sklearn to some extent. the ecosystem for web development is completely different.

[weak mortar]

nice to know that i have to choose my path then
before digging deeper in python

[valid trail]

You could try some game development. It's fun and a good intro to control flow and logic

pygame, tkinter for those

[oak turtle]

I would like to get involved into the data sciences field, any recommendation?

[stone juniper]

try #career-advice

[oak turtle]

Thanks

[misty dirge]

@weak mortar I had to remove your messages, as they were off-topic.

[fallow terrace]

@weak mortar this channel is for the discussion of teaching programming. If you want to use different commands, please use #bot-commands.

[muted scarab]

Do you think CTF type exercises are applicable for learning different aspects in programming and software development?

[misty dirge]

what are CFT type exercises?

[muted scarab]

Capture the flag

[misty dirge]

how could capture the flag and programming be related?

[oak breach]

it's a cybersec exercise where you have to find hidden security flaws

[muted scarab]

As hsp said , it's a very common hacking exercise and competition.
It's pretty successful with cyber students and even in high schools.

[misty dirge]

@zenith smelt @rotund belfry this channel is for discussing the methods and practices of teaching. You're welcome to participate, but please respond to a message or introduce a discussion topic.

[zenith smelt]

i understand, I'll try to interact as much as I can.

[pure fog]

I don't think they're very relevant outside of security specifically

but yes i think anything that forces you to get "hands on" and write real problem-solving code, in a less-controlled environment, is good for really learning to program

[stone juniper]

this channel is for discussion about how to teach computer science, but you can open your own help channel from #❓｜how-to-get-help

[bitter sinew]

well in my case the first steps is making a terrain to test the IA its mean to be some "place" where the IA can learn im using genetic algoritm to get the succes of a task (follow a ball)

[subtle owl]

Thanks!
But I am not sure to follow how that relates to the question.
(still sounds interesting though as I am also interested in genetic algorithms in general)

[weak mortar]

Not Python-specific, but I've got a programming question.

I learned Python and Golang over the past 5 months, and I've build some intermediate things like a Flappy Bird game and an AI to play it, a sorting algorithm visualizer, and some CLI tools. Lately I've wanted to take my skills to the next level and make things like emulators, libraries, and maybe even a programming language (These would be in Golang of course). But I don't know how to move on from being intermediate. I don't know what resources I need or what I need to practice and learn. What should I do?

[pure fog]

note that this channel is about teaching programming.

but to address the question, you can start practicing library development by splitting your existing applications into "library + frontend". so your cli tools could be refactored as some library that implements whatever the tool's functionality is, with a cli "shell" around it. and you'd refactor your tests accordingly. you can also try to contribute to established open-source projects, usually they have several small bugs/features that are good for first-time contributors to work on

[weak mortar]

Sorry, didn't know what channel to put this in

Have any good open source projects for new contributers to start with?

[pure fog]

maybe pick a project that's already related to a field that you are interested in. or maybe even a library you used in your past projects

admittedly i don't know what channel this would go in either. #python-discussion moves too fast

maybe we can justify it in that you are asking a question about "how to teach (myself) programming beyond the intermediate level?" 🙂

[weak mortar]

I'll start looking for some open-source projects to contribute to. Thank you!

[indigo pendant]

I have Python in my class as a computer scientist and I am failing.
Now I will have holiday of 2 weeks and need to work on my python skills.
How did you guys got good with python?
Watching tutorials or combining it with a own project?

[upper flint]

I think that building out a project is one of the best ways to learn, since it combines the "theory" of reading documentation / tutorials / lessons with the "practice" of writing / debugging / running the code

[indigo pendant]

Should I combine it with a tutorial or should I just head to building a project and keep researching on my way about the code I want to implement?

[upper flint]

Try both! Following a tutorial is a good guide, but then try to do the same thing from scratch, without using the tutorial (only documentation / implementation details you find on your own)

[pure fog]

imo reading documentation is "practice" and is a skill that needs to be learned. same with deciphering error messages and debugging

[upper flint]

Yeah thats a good point. I guess my separation was split along reading / writing code which isn't the same as learning / practicing

[pure fog]

yeah valid

[weak mortar]

Hey guys, this program enables me to create a Customers Table and a Products Table. How can I implement a "Buy" function that makes the user select a Client then Products and sends this data to a "Total Sales" table ? I'm using Tkinter and MySQL

Code here : https://pastebin.com/bpciapwe any help would be amazing 🙂

[wooden grove]

@weak mortar #databases would be the appropriate channel to ask in.

Or maybe #user-interfaces. You could also claim a help channel. See #❓｜how-to-get-help for instructions on claiming a help channel.

[misty dirge]

Topic: How do you explain the concept of state? I developed my understanding of it gradually, and I think if I were to try to explain it to my former self, I would have misunderstood the explanation but thought that I hadn't.

[vale current]

well, I guess I would call state as sort of the "memory" or "condition" of something. Methods on a class have state, since they can function differently depending on different data set on its attributes

[misty dirge]

String methods (or any method free of randomization and involving only immutable types) are stateless, though

[vale current]

hm, thats true

[tidal orchid]

if I had to come up with an example from scratch (unrelated to a problem they are already trying to solve) I would probably just use a list or typical RPG stuff (such as your character's items) for example

state never really seemed to be a difficult thing for people to pick up though? at least from what I've seem

possibly a hot take I guess:
it is fine to misunderstand it conceptually at first as long as you understand it well enough to make use of it
*and make sense of what's happening in your program

[orchid thistle]

i think having an understanding that's "good enough" but actually missing something significant has the potential to cause problems in situations where that understanding isn't good enough

although i guess you can't dive right into how names and values work before teaching literally anything at all so you do need a certain amount of simplified models in order to teach things in an order that makes sense

[orchid thistle]

especially since a lot of people probably wouldn't think to go back and reinspect their models of the basics of how python works

[pure heron]

I've never had to try to explain state before. This is tougher than I expected. Here's a swing at it: your program's state is every piece of information that could change the outcome of a future operation.

[stone juniper]

alright, first imagine a tape extending forever in both directions...

[noble aurora]

I've never actually explicitly learned what "state" was, it's kinda just everywhere in languages like python, so you have to know sorta what it is

[stone juniper]

seriously though, those little robot programming games are a good way to grok state

i guess i wouldn't even try to teach state in the context of python

[noble aurora]

wdym?

[stone juniper]

to introduce the concept, i mean

beginners already have trouble seeing the semantics behind the syntax, i think just imagining a simple machine is a much more intuitive way of understanding state than looking at source code and from there it should be fairly easy to extend the idea to any procedure

[pure heron]

even that doesn't really cover it, I think... That covers memory, but memory isn't the sum total of a program's state

[stone juniper]

that was just a joke about Turing machines, because you have to explicitly define their states

[pure heron]

oh - I thought it was glib but semi-serious, since that's an adequate (if not ideal) way of explaining the state of an algorithm

but it doesn't scale up to a full networked program running on a multitasking desktop OS

in particular, it doesn't scale up to more than one thread

[stone juniper]

it kind of does

1. the machine itself can be in various states
2. when you think about it, the symbols on the tape and the current position of the machine also determine its behavior
3. now extend that to everything being kept track of everywhere that might affect the program

[pure heron]

I suppose tapes could cover memory even across processes, but a Turing machine only has one instruction pointer, and you need multiple instruction pointers when you talk about parallelism

[stone juniper]

well that's just a great opportunity to show them a bijection from the naturals to finite tuples of the naturals

[pure heron]

unless you extend your model to include the operating system and its scheduler, and the other machine, and its operating system, and its scheduler, I guess.

[misty dirge]

You mean like 1 -> (1,)?

[silent goblet]

I think he meant that tuple[int, int] is isomorphic to int and such

[noble aurora]

wait how

[noble aurora]

wild

[cosmic aspen]

Lmao like this I hab code in my Python lib

[oak breach]

ok but tuple[Real, Real] -> Real is wilder

[stone juniper]

@noble aurora and you can apply this multiple times to show that you can get a bijection to any finite length of tuples, and once more to show that you can get all of them at once

[noble aurora]

right

[stone juniper]

which is just to say that multiple states are just one state in disguise

[noble aurora]

naturally

[tough hollow]

can we make threads for different topics so I can see where the conversation came from?

[misty dirge]

the thread rollout is something we've been discussing internally. I'll bring this suggestion to the rest of the staff.

[tough hollow]

ok!

[misty dirge]

@wheat schooner @maiden tree this is a discussion channel; please conduct help in a help channel (see #❓｜how-to-get-help) or a topical channel (like #databases)

[waxen fox]

would you all say teaching name mangling as a way to make attributes "private" is a good thing in order to avoid confusion?

for example, in my python class, we were learning about _ vs __ for attribute name prefixes, and the teacher said that __ was better for making attributes "private", even though most people use _ for signifying that an attribute/variable shouldn't be touched and __ is mostly used for preventing inheritance collision

[misty dirge]

I would only teach a single leading underscore for "private" attributes/methods. the effect of name mangling is only realized if you make a subclass, and even if you are, the need for name mangling is only present if you plan to have a particularly complicated class hierarchy. And at that point, you should probably just use java.

(usually I wouldn't say that someone should use java, but if someone insists on having such obfuscated behavior, I don't want them adding it to our ecosystem 😄 )

[pure heron]

There's plenty of deep class hierarchies in Python, like what the io module has (and therefore what open returns)

[misty dirge]

lest we get into a debate about whether or not name mangling is a good language feature, I don't think it's worth going into for beginners.

[pure heron]

Double leading underscore name mangling is only useful in one really narrow circumstance:

1. You own a class, and
2. There are subclasses of that class that you don't own and don't know the privates of, and
3. You want to add new private attribute or a new private method to your base class without it colliding with a private name chosen by an existing subclass

[misty dirge]

Would you agree, then, that double leading underscores should not be covered in intro to programming curriculum?

(changed from only to not because I changed what came after it and didn't fix the whole message )

[pure heron]

I wouldn't teach it to beginners, no. Point 2 in particular means that name mangling is only useful for library code. For application code, you know all the subclasses that exist, and can check what privates they have and choose a name that doesn't already exist in any subclass.

[misty dirge]

@waxen fox are you satisfied in our treatment of your question?

I like your point about its potential usefulness in library code.

[waxen fox]

yep!

[ripe folio]

I agree, I hate seeing name mangling used anywhere. They have one user, which is to handle two classes needing to name the same attribute differently without collision.

Sadly, I've seen it used where they should've been internal.

The less people who know about name mangling, the better.

[misty dirge]

There is no name mangling in ba sing se.

[weak mortar]

Any hints on where or how to Dive in into python source wide?

I mean, i fought my way alone with SQL Spend Like hundrets of hours because i Had a Task ans Goal in my Profession. I reached the Goal

Same with SAS, Goal, Task, was existing, so i worked ans worked on it

[waxen fox]

agreed; I'm not sure why they chose two leading underscores for name mangling either, given one leading underscore was used for showing that an attribute/variable is protected, people would assume since it mangles the name and has leading underscores, it makes attributes private.

[misty dirge]

you're saying that these people think there are specific protected and private access modifiers that are language features in Python? because "protected" is a whole separate thing in Java.

[waxen fox]

more so that people will assume the mangling is for making a variable private because both the convention PEP8 defines for attributes that shouldn't be touched but can technically be touched (single leading underscores) and name mangling use leading underscores, sort of categorizing them into one

idk if I'm making sense tho 🥴

[misty dirge]

protected variables are a specific thing in Java that doesn't have an analogue in Python's psuedo-private system.

[waxen fox]

i thought they use the term "protected" in PEP8 for single leading underscores?

[misty dirge]

maybe. perhaps we should make a PEP to make the nomenclature more consistent: PEP 9002.

[waxen fox]

a PEP for the PEP

[green warren]

The lack of access specifiers is a serious annoyance I have with the language

[fallow terrace]

update: I came across this explanation in a tutorial, and it makes a lot of sense.

A variable that you can’t vary doesn’t make much sense
Why not? The name is taken frome mathematics, where it indicates something immutable and rebindable – just the same as your ‘bindings’.
The ‘vary’ comes from the fact that a variable may refer to an expression of varying or indeterminate value, like  y = f(x).

[pure fog]

"variables are labels attached to data"

the name is entirely vestigial and borrowed from math

i suppose that learning programming was easier for me because i did know math, so i already had intuition for variables and functions that wasn't hard to adapt

[pure fog]

this is very good too

it "varies" because you don't necessarily know what it points to

also it's worth considering the difference between an "immutable variable" ("immutable binding") and an "immutable data structure"

and yes i am big -1 on teaching double underscore attributes

heck i am -1 on teaching private/protected attributes at all. it's the kind of thing where if you teach it too early it becomes a point of confusion and unnecessary cognitive noise, but if you teach it later it's almost trivial

people who teach "getter/setter" OOP in python are really doing their students a disservice

[green warren]

I'm not really a fan of getters/setters either. If a user can freely get/set a member variable via trivial getter/setter methods, then that member variable is not involved in any class invariants, and may as well be public.

[misty dirge]

What do you mean by class invariant?

[green warren]

I mean, if it's important how the value of that member relates to anything else about the class's state.

I'm beginning to think that "classes" are a kind of inherently confusing idea anyway, and that one should really focus on interfaces. The reason you bundle a bunch of stuff together into a class is because you want to do something, and you want to abstract that process away and create something out of it that is more convenient to use and think about.

And then the question of "how to access data members" becomes a bit nonsensical, because that was never the goal in the first place. You provide inputs and receive outputs.

[misty dirge]

as much as I hate arguments from compromise, I think it really is a case of "it depends". Classes make a lot of sense in the case of discord.py, and many of the classes there have both methods and attributes.

[pure fog]

this is the "original" intent of OO programming: all communication between objects is a "message" that is "passed"

"give me your value for x" is as much a message as "compute this thing please, with these inputs"

so in a language where that design is carried out, then of course you need getters and setters, because the notion of a public attribute doesn't even exist

in python it's completely the opposite: objects are basically glorified hash tables, and methods are just funky single-dispatch functions. so in that context, it makes total sense to have publicly-accessible attributes, and it makes no sense to have getters and setters. = is a getter and setter

[winter spear]

I personally like getters and setters in the form of properties purely for documentation

I wish data descriptors showed type hints like methods

[pure fog]

well you can type-hint instance and class attributes, that's what i do

i agree that support for data descriptors would be really really valuable (and it's frustrating that it's not a thing)

[winter spear]

I feel like I bring this up with you a lot, but it doesn't show up in help() feel like I'm like one of few people to use it. Probably bc of my job

[pure fog]

i don't remember talking about this (different username?) but i agree with you that it should be better in both of those ways

@green warren @misty dirge in general, i like to say that classes have 3 purposes. i used to say 2, but i thought of a 3rd one:

1. grouping related pieces of data together, e.g. an AppSettings class that contains all the configuration data for your app, or a Point3D cartesian coordinate in a simulation

2. encapsulating local mutable state, e.g. storing login information in a discord client, or some handle to an external resource like a file or network socket, or implementing a state machine (this one can be done without classes too, but it's a very different programming style)

3. implementing single-dispatch polymorphism in languages that don't have function overloading, allowing end-users to customize behavior directly or providing them with hooks that allow them to easily do so (e.g. by subclassing or implementing a protocol/interface)

[pure fog]

how often do you use data descriptors and help()? mucking through unfamiliar django codebases?

[winter spear]

I employ data descriptors myself regularly

I'm a big fan of descriptor protocol in general. Recents have been a lazy property that works with __slots__. functools.cached_property doesn't. Another is an instance state based dispatching system for methods

@pure fog here is an example of state based dispatch I shared on reddit https://gist.github.com/Delengowski/48f3aa87536caaaf037c4f1116f85dfc

[pure fog]

cool, not something i personally need but i see why you'd want it

apparently (as i just learned in #type-hinting ), mypy does support descriptors now, even if you can't explicitly annotate them yet

so you can annotate the return value from __get__ and mypy has at least some support for using that

[dreamy sparrow]

I know that the PEP8 reccomended formatting for python variables is to use snake format (eg. my_var). But a lot of the code I read for integration/data-science projects uses variables in camelCase. Is this actually a problem, should I just do a quick “replace all” and change the variable names, or is the effect negligible.

[pure heron]

It has the effect of making the author look inexperienced. If someone writes code that doesn't follow extremely common Python idioms, it strongly suggests they haven't read enough Python code to actually become familiar with the language.

[dreamy sparrow]

Yeah that’s fair enough, it’s strange because the files I usually find this error in is made by people who clearly have a very good proficiency. I would agree that conventions are essential in python for readability and integration purposes. It’s definitely something that needs to be emphasised in the learning stage

[green warren]

camelCase (with initial lowercase) looks so awful, I don't use it for anything in any language. snake_case for variables and methods and namespaces, UpperCamelCase for the names of classes.

[ripe folio]

Support from who?

[misty dirge]

a lot of data science code is egregious, both stylistically and comprehensibility.

[native hearth]

But have you seen physicist code?

[green warren]

Oh yes. Many of them go on to become data scientists after they learn to code better

[pure fog]

from type checkers. but apparently it has more support than i realized

[pure fog]

this, bigtime. not following the style guide in any language is a sign that the programmer isn't interested in writing idiomatic code, it's a smell for "this is going to hurt" when reading it (and possibly a warning sign for bad/weird/unconventional API designs)

[rotund vortex]

lol when I was interviewing for my job they asked me to write some pretty basic code on whiteboard. As the guy revealed me later they were afraid that my code would be completely incomprehensible and they wanted to check that I am able to solve a trivial issue with similarly trivial and clear code

very often in physics you don't care about how the code is written you only care that it runs few times. Also the very common pitfall is this: oh this will be a one-off code, I'll never use it again so why bother

(I think that's a common pitfall all around though, not just physicists specific)

[pure fog]

this is a nontrivial issue in hiring and yes you need to check for it. you can't have people writing "scientist code"

[rotund vortex]

yeah and I think what is even more difficult is to decide whether this person can learn on the job to write proper code or not..

[native hearth]

if they've gone through learning all the fancy shmancy sciency things, a little bit of coding style and convention cant be that hard

[pure fog]

depends on how stubborn they are, and how strong your leadership is

[rotund vortex]

exactly. But the question is, whether they are willing or not. I think anyone can learn that but it might require some effort on the onboarding side to convey the why... The mindset should switch from the one you had in academic env to a more practical one, and that could get tricky

[pure fog]

it depends a lot on whether the engineering leadership is willing to tell someone with 10 years exp and a phd "no, and i mean no"

if you aren't comfortable enforcing code style conventions then don't hire people who write bad code. otherwise you can hire them but you have to make it very clear that they are expected to improve their code quality up to some kind of basic industry standard (not that industry standard is any good anyway)

[rotund vortex]

i.e. in my PhD years I probably wasted close to 1 million cpu hours and I had no reservations or regrets about that cause it's something "free of cost". And the mindset was "well, let's run this and if it crashes I'll deal with it later". Not something I'd ever do now I think 🙂

[pure fog]

oof

for what it's worth, i don't quite understand how scientists/researchers got in that mindset. even when i was first learning to program in school (for social science), i didn't understand why people were so resistant to writing readable code

it always seemed like this kind of "punchcard cosplay", like if your code was hard to read it meant that you were the academic equivalent of a 1337 h4x0r because that's how the OGs did it

take some goddamn pride in your work. i too make plenty of compromises under time pressure and knowing that my code is "disposable", but there's a limit below which you're just making a huge mess for no extra benefit

[rotund vortex]

lot of them (us, in a way 😁 ) are told stuff like "you are a researcher, not a programmer so you must focus on producing new and interesting results, your phd is in physics and not in writing code". that kind of things

[pure fog]

sure, but like... you need the code to do the physics. it makes no sense to me

[rotund vortex]

I might be exaggerating here and oversimplifying but that's the gist of it.

[pure fog]

no i know that's actually how people think. "it does't matter" + "i can't be bothered"

i just never understood it

[green warren]

The flipside of this is that nobody in science publishes their code, only the results they got from running the code.

The code does not earn them anything toward their next academic job. Publications do.

So it's not just that somebody can't be bothered, but the entire incentive structure encourages you to write quick and dirty code that gets the job done, and move on.

I haven't given anyone a literal whiteboard coding test, but I do ask them questions about coding practices and structures.

[pure fog]

quick and dirty code is buggy code. the problem is really that academia encourages quantity over correctness/quality

and you should be required to publish your code, data, etc. or at least make them available for inspection (eg. you can't really "publish" LHC data)

[green warren]

Large collaborations like the ones looking at LHC data probably have pretty decent code, actually. "Scientist code" is really more of a problem with one-off projects

[pure fog]

true. my point was just that sometimes the data is really huge

[rotund vortex]

Some do publish though. Like I was working and making contributions to Quantum Espresso (you can find main repo on gitlab). And it's big enough so that me saying I contributed has pretty much 0 doxing risk 😂
But many won't publish. Or will publish later when they did what they wanted as monopolist

We also usually publish calculation parameters and sometimes even straight up input files. So it's not that bad everywhere

[pure fog]

quantum espresso is like when i make espresso at home, and one shot comes out great, and the next shot tastes like shit even though i used the exact same recipe?

[rotund vortex]

Yes

[green warren]

I literally wrote a computer science algorithm paper and didn't publish the code. The published work has only pseudocode.

[misty dirge]

why not? 😠

[green warren]

I didn't think it was well written, and wanted the paper to focus on the actual algorithm

[misty dirge]

my work often involves reproducing the results of papers. though I've found that often, even when there's code for it on github, I'm no better able to comprehend it than from the specification in the paper.

[green warren]

Yeah, I did actually include a link to it on GitHub, it just doesn't literally appear in the journal

And it's written in Mathematica (or I guess they're calling it Wolfram Language now)

[acoustic prawn]

Good evening

Can someone explain to me how a recursive function works?, question in the help-burrito Help Channel

[jagged heron]

hello there nice people :) i just wakted to ask of your oppinion. what do you think. should i remove a person from github project if he is no longer working on it ?

[jagged heron]

programming a n! function in recrusive way. pretty sure you will get the concept of recrusive function when finishing the excersize :)

[pure heron]

This channel is for conversations about how to teach programming. Try asking in #python-discussion

[subtle owl]

I found myself digging into the source code written for papers or wishing there was such code.
Sometimes, there are some ambiguities the authors haven't thought about.

Also there is no such thing as shame for the code, especially when it comes to papers

[misty dirge]

how recent is that name change? one of my coworkers likes to use mathematica

[green warren]

The name was changed a few years ago

[misty dirge]

that explains it

[pure fog]

i imagine that most people still call it mathematica

[green warren]

Everyone still calls it Mathematica

[misty dirge]

I'm gonna call it cupcake

[green warren]

Cupcake Language

[stone juniper]

clang for short

[cosmic aspen]

What it changed its name? Lmao

[rotund vortex]

Wolfram Language

[neat canopy]

what is a good project to help teach a low intermediate student

[wispy comet]

I think the appropriate project depends on what functionality you want them to learn

[subtle owl]

I try to look for projects which are basically about what the person knows + skills I want then to acquire.
That enables them to not have to deal with too many unknown at once and to build on what they have learned so far

[sage jasper]

do you have some go-to projects?

[subtle owl]

I don't have any specific go-to project in mind.
In general, the mentee would either already be passionate about something or it would happen in the context of work where the projects are already set.

I would recommend to look at what is the person passionate about first, and then looking at a clone of a popular product by default.

[shell mesa]

Is it ok at ask questions as a student here?

[subtle owl]

The solution is not to ask to ask, but just to ask. Someone who is idling on the channel and only every now and then glances what's going on is unlikely to answer to your "asking to ask" question, but your actual problem description may pique their interest and get them to answer. 

[shell mesa]

I wanted to ask...

I want to be a proper programmer. But currently I am just writing code that works.

It's like the difference between knowing how to cook vs being a proper chef

I want to be a proper chef

(I forgot to send)

[subtle owl]

sure, so what's the question? What's stopping you?

[shell mesa]

What's the path, how do I get started in being a proper programmer (and not someone who just knows how to write code)?

(I already know intermediate python)

[subtle owl]

There is no specific path. It's part of the craftmanship.
Learn about the patterns, the language, when to apply which technique or not and practice

Also confront ideas and keep an open mind as people come from different background and had to deal with different constraints

[shell mesa]

What is Patterns?

What I need to learn more:
The language (tool)
Techniques (mindset?)

[subtle owl]

There also many books about improving the craftmanship from different angles. From "clean code" to "righint software" to "design patterns" from the GoF to "the art of scalability".
And that's not even talking about the leadership related books

[subtle owl]

patterns as in "design patterns"

[subtle owl]

https://www.amazon.com/Head-First-Design-Patterns-Object-Oriented-dp-149207800X/dp/149207800X is a good intro to design patterns

[shell mesa]

Ahh thanks

[subtle owl]

https://www.amazon.com/Design-Patterns-Elements-Reusable-Object-Oriented/dp/0201633612 is the classic that I still use as a reference

[shell mesa]

So when making programs I have to think like a programmer, I have to apply techniques and practices

[shell mesa]

What questions should I ask myself to know that "I know and understand my software that I made (website, api or anything)"?

[subtle owl]

Are you able to explain it to someone else?

[shell mesa]

I can explain the general idea.

But in-depth I can't

[subtle owl]

being able to explain and teach it is a good test for whether or not you understand something

[shell mesa]

But I if I don't want to rely on other people? I don't have friends and getting decent and actually helpful feedback on discord is kinda hard

[subtle owl]

you don't need other people.
Think about it like writing a blog post (or two) about explaining it

It forces you to think about the concepts you need to go over, the ideas, the best way to go about it and how they fit together.
Personally, I have a private blog to write down my ideas

[shell mesa]

Thanks!

[languid yacht]

Is the advice "just keep writing code" really helpful I keep getting this in response to "how do I get better"

[native hearth]

"how do i get better" is a really vague question, "practice" is a really vague answer

Obviously if you keep writing bad code you wont get better, you should be more detailed in your questions if you want more targeted answers

[languid yacht]

Then what should I know as a backend developer who will use Django as a backend framework

[noble aurora]

presumably, django

[native hearth]

Thats not any more detailed than "how do i get better" tbh and answering that goes beyond the scope of the channel

[misty dirge]

It's like the difference between knowing how to cook vs being a proper chef
The difference is if someone is paying you to do it as a career. There isn't necessarily underlying knowledge that fundamentally legitimizes your programming ability.

[vale belfry]

How do athletes get better? How does a mathematician get better? How do musicians get better? Practice.

[subtle owl]

Think more about deliberate practice.
For instance, if you have some problems with DP, then you want to practice more DP. Practicing randomly would help you with DP only by luck if it relates to it.
Similarly, if you want to learn more about packaging your software, then practicing packaging your projects would yield more results than doing some leetcode

And one important thing is to also setup the criteria for success. "Getting better" can means a lot of different things depending on who you ask and where they are in their journey

[vagrant crater]

I personnaly get on personnal projects, I try to make stuff that university wouldnt' ask me to make and I basically discover new aspect of programming every day

[misty dirge]

@weak mortar your comment was off-topic for this channel. See #networks

[fallow terrace]

@eager coral This is not the right channel for that. #python-discussion or off topic would be better places.

[pure fog]

i think so, although i think it should be "keep writing code, and keep reading other people's code"

[bleak drum]

I regret not taking my undergraduate programming classes more seriously and am definitely paying the price now at my current internship. I have been reviewing concepts and doing practice coding tasks while working on my current project but I feel as if I am never going to be as good had I taken my courses more seriously. What can I do to make this process more bearable? EDIT: sorry if this isn't the right channel to post this, leave a reply and I'll remove from the channel thanks

[weak mortar]

its ok. Definitely study the recommended textbooks from ur programming classes and fill the gaps in your knowledge. Rewatch lectures as well. I kind of dislike when students say cs theory is not important for industry which is simply not true

[misty dirge]

@bleak drum if this is your first time writing large programs outside of your assignments, it might be that you'd have ended up feeling caught off guard regardless. Programming in real world circumstances isn't always the same as school assignments that have a narrow learning objective

How much longer does the internship go on for? And how far into the curriculum are you?

[split ice]

I have a question yall. What do you really need to know, to "know" a programming language?
Cause of course your not going to go super-advanced in all languages.

ping me @split ice

[pure fog]

it's not about what you know, it's about how deeply ingrained in your brain it is

it's a matter of achieving fluency with syntax, idioms, and standard libraries

[pure fog]

get a good pen and a notebook and take a lot of notes

i'm serious. draw flowcharts and activity diagrams. write out pseudocode. put line numbers and write your comments on those line numbers

if you're really struggling, drop into the debugger and poke around to see the current state of the application. and of course do not afraid to ask for help, that's what internships are for

[misty dirge]

You can say that you "know" a programming language if you can write code in it without having to refer to a reference for how the language itself works. If you don't know how a given expression is being evaluated, you probably don't really "know" that language beyond a cursory understanding of how programming languages work in general.

[split ice]

fair enough

But of course you have to still search up some syntax here and there of course

your not going to know everything, that goes for anyone

[split ice]

ty

[pure fog]

syntax? probably not, unless you're switching between languages

i can't remember the last time i had to look up python syntax

[split ice]

What I mean by that is let’s say your trying to figure out something in python for example that doesn’t exist that you don’t know about

So of course your going to search it up

Your not going to know every little piece and syntax in python (ex)

Manifestly, of course

I’m referring to advanced syntax that is not beginner based

Something that doesn’t just come off the top of your head

[desert patio]

i think i know every piece of syntax of python now

star unpacking in subscripts, starred try{-except}[-finally]/try-finally

and this ```py

[a, *[b, c], d] = 1, 2, 3, 4
a, b, c, d
(1, 2, 3, 4)

and this ```py
for i in 1,2,3:
print(i)

[cerulean nebula]

python does leave room for some unintuitive expressions though, this tripped me up a bit when I first saw it ```py
my_list = [True, False]
print(1 == 1 in my_list)
print(2 == 2 in my_list)

[wise basalt]

can someone help me with a school exercise from python

it's in dutch tho

[subtle owl]

Hi and welcome aboard!
However your question has nothing to do with the topic of this channel. You may want to check #❓｜how-to-get-help to increase the odds of getting some help

[subtle owl]

Note also that we won't help you cheat your exams or tests

[wise basalt]

it's not a test

[subtle owl]

just saying it in case. I can't read dutch

[wise basalt]

oh oke

[weak mortar]

heyy i was wondering anyone know if u rlly need a diploma from school to be a software engineer u can self teach too right

[misty dirge]

try #career-advice

[misty dirge]

if you're a young person with no professional experience, the self-taught route isn't really a viable option for most types of developer positions. there's no way for potential employers to verify that a self-taught person actually knows what they claim to know unless they have tangible experience doing it, but tangible experience is difficult to obtain without a degree.

a computer science degree will pay for itself, so if you think you're up to being a developer and can get student loans, it's a worthwhile investment.

[weak mortar]

what about online courses

[misty dirge]

if they're from an accredited institution, that's fine, though they won't carry as much weight as a degree.

[weak mortar]

true but a company would hire someone without a degree but can show hes rlly good then some dum dum who got a degree but is pretty bad

[misty dirge]

"a company might hire someone with a degree who is incompetent, therefore a degree will not necessarily increase my chances of getting hired" -- is that what you're saying?

[weak mortar]

not exactly more like if u can somehow show ur experience that ur just as good as someone with an degree maby better u will have equal chances

[misty dirge]

if you can show that you have experience, there are companies that will accept that without a degree. but it's usually only professional experience that companies will recognize. that is to say, it's very difficult to break in without a degree.

[pure fog]

the thing is that the probability of someone with a relevant degree being incompetent is low, and the probability of someone without a degree begin incompetent is higher

economists call this "signaling" - it's very difficult to observe actual skill on the job market, so candidates must signal their skill with easier-to-observe attributes like "i have a degree in CS"

so you are working against the fact that you lack a strong signal that the market has come to expect, and you must find a way to make up for it. the signal is very strong and due to automated resume scanning it might be difficult to get past it. but i think this is #career-advice talk, no?

[misty dirge]

Those with CS degrees (and anyone else who wants to participate), how did you feel about how the curriculum was broken down into courses? Were there any courses that you feel were too narrow in scope, or which were too broad in scope? Did the order in which you were required to take the courses make sense?

I thought the theory of computation course that I took was pretty weird. a lot of what we learned (like the pumping lemma or cantor's diagonal method) I felt was just there to fill up the time so that it could stand alone as a course. and then the programming language theory course that came later didn't do very much to relate that material to the theory of computation course, even though the two are pretty interrelated.

[weak mortar]

lot of the courses i took were too board in scope. like there's just too much to cover even the upper level cs courses. only few courses like OS, DSA we dove deep. There is just not enough time so it felt like surface level knowledge even tho the courses were tough

[misty dirge]

Interesting. What was the title of a course that you felt was especially broad, and what was covered in it?

[weak mortar]

I found computer networks (iso model, protocols, network security etc) and programming langs course too broad. for programming langs stuff that was covered was (comparison bw imperative, oop, functional, and logic langs, control structs, binding, scoping) basically formal description of all types of comp languages.

but ig thats what undergrad education is for. to give a good overall foundation for cs. if u wanna learn deeply abt specific area then u can do masters or phd

[summer helm]

Did you have to relearn courses that you already knew?

[zinc wedge]

Is this the channel for learners?

[fallow terrace]

no, this channel is for talking about teaching programming

[arctic wren]

I took an advanced programming class that pretty much made sure we understoof classes to a decent extent in C++ in the first half after a semseter of super basic python, basic java, and a semester of basic data structures. The other half of the advanced programming class was just projects that had to use 3 or more languages for the most part. Learned a lot in that class

but most classes are what you make of them from my experience

[jagged wolf]

do i ask for ask for help in this server

[lunar wave]

You can. #❓｜how-to-get-help explains our help rooms.

[tired quiver]

Can anyone recommend a solid way of learning/growing with Python? I got introduced to programming through some university courses and had a pretty fun time. Now that I'm no longer in those courses, I am looking for a way to become stronger, knowledgeable, confident in Python to the point of finding a job/career with it. I downloaded some apps (DataCamp, Codecademy Go, Mimo), I know of a few YT tutorials and websites, but the experience feels shallow. I guess I need to start working on projects to test my understanding, creativity, and problem solving? (I'm not saying I know everything at this point, far from it.)

[misty dirge]

Learn by doing. Once you've learned something, you should do something to apply what you've learned. Otherwise, the information probably won't stick.

It's also worth noting that there's a point at which you're no longer learning Python itself, but learning how to do certain things, and learning how to use tools that happen to exist in Python to accomplish those things.

[main bane]

i am also a beginner and i do that after every new thing i learn i make few projects combained with the other things i have learned before and it really helps

[tired quiver]

I'm just so confused as to what to do and what to believe in. Do this, do that, don't do that, this won't work, first you need this... I wish there was a more straightforward path. I feel like there's too much risk, too much investment for an iffy outcome. Is it just me?

[lunar wave]

What risk is there that you see?

[misty dirge]

Sounds like this should move to #career-advice

[cerulean bough]

@tired quiver

I'm just so confused as to what to do and what to believe in. Do this, do that, don't do that, this won't work, first you need this
Dont't just do what somebody (or some tutorial) tells you to do, try to understand why you need to do that.
Do this (not only in programming, but in everyday life) and you'll know what to do on your own.

[misty dirge]

@tribal breach see #data-science-and-ml

[weak mortar]

Hello, Python community. How would you recommend an absolute beginner start to learn Python? I have familiarity with Linux and a working knowledge of how computers function, but other than that, I'm an absolute beginner. Where should I begin?

[misty dirge]

!resources

[worldly dew]

Resources

The Resources page on our website contains a list of hand-selected learning resources that we regularly recommend to both beginners and experts.

[misty dirge]

Keep in mind that this channel is for discussing the methods and practices of teaching.

[ionic lagoon]

I'm creating this project https://github.com/jwlei/YOLOv5_DeerDetection/tree/master/DeerDetection

And I was just wondering if I actually write readable code, I've been trying to format it as best I could

[vale current]

Consider using a linter, e.g flake8

it'll point out any code that might not follow pep8, as well as other conventions and the like

[candid tide]

Hi everyone, I’m new here and to programming and would be happy if someone could help me out from the very beginning

[wooden grove]

@ionic lagoon That's not really on topic for this channel. Perhaps claim a help channel: #❓｜how-to-get-help

@candid tide Also consider claiming a help channel or asking in #python-discussion

.topic

[fleet apex]

**What are the most common difficulties/misconceptions students encounter while learning to program?**

Suggest more topics here!

[lunar wave]

randomly swings in That the name of an object/variable has anything at all to do with the contents!

[misty dirge]

relatedly, I think beginners tend to conflate objects and variables. I encountered someone recently who thought that creating a class created a "new type of variable". I've also encountered people who thought that creating additional variables for the same object creates copies.

[lunar wave]

It's probably one of the largest obstacles that I see someone have to get over in early programming. One of my teammates still struggles with the abstract concept that just because he named something a list, that it doesn't assert or promise that's what is in the variable.

I wonder if some of the challenge comes from the dichotomy of "a program only does what you tell it" and "don't blindly trust what's in that variable you created".

(clarity that this teammate in question is an app engineer, a product owner mindset, and not an experience coder)

[misty dirge]

I was about to ask, because it's disconcerting that someone with that misconception would have made it that far without having been corrected.

do they have experience with a statically typed language with compile time checking?

[lunar wave]

He is leveling up wonderfully well considering he threw himself into the deep end without checking the current. He's mind is sharp, he spots bugs in code on an end-user level long before I do. Programming concepts though seem to just trip him up hard, especially around variables and their contents.

[lunar wave]

He has far more experience in networking, network typography, and network operations. To my knowledge, he had never done more than a few simple powershell scripts before offering to work on smaller automation stories for us.

[desert patio]

python doesn't copy unless you explicitly tell it so

[stone juniper]

Beginners tend to focus on the syntax of the program:

"What should I write?"
"Do I need to add a new variable?"
"What syntax do I need to solve this?"
when they should be thinking about the semantics, since those are the terms in which the problem is solved.
It is understandable, given that the syntax is what they see in front of them and what they do when coding. I think that making them write pseudocode helps train them to see the semantics behind what they're writing. Another is giving them incomplete solutions so they see that the right answer is not a magic key that solves the problem; it's built of steps which operate on the objects created in previous steps.

...or thinking that the difference between
["a", "b", "c"] and "".join(["a", "b", "c"]) is the brackets you see when you print them

[misty dirge]

someone who isn't me (but I don't remember who) pointed out that beginner assignments are centered on taking inputs and printing out the right thing, and that this might be why beginners tend to conceptualize everything as strings (even if they don't think of it that way) rather than as data.

[stone juniper]

hm that might've been me lol

[misty dirge]

then you certainly deserve a cookie. 🍪

[desert patio]

that somehow isn't me when i began python

probably because i first learned lua from roblox
roblox lua tutorials aren't focused on input-output, they're more focused on syntax and memory (storage)

[normal comet]

Is it possible to draw sketches without using the sketchpy lib class?

[stone juniper]

this channel is for discussion about how to teach computer science, but you can open your own help channel from #❓｜how-to-get-help

[waxen fox]

this one isn't really too big of a deal but i suppose it could still count as a misconception -- that int, str, range, etc are all functions

[vale current]

Perhaps return VS print?

[white bough]

Everythings an object,

I would say that in a python context int, str are classes and range is something I would describe as a function.

If we were pedantic we can say that they are all functions as they are written in C and therefore cant be classes.

But yeah, the repr of int,str will tell you they are types and the repr of range should tell you it is a builtin.

[vale current]

Could you argue that anything that is callable is a function?

[neon pasture]

range is also a type

[vale current]

Avery was making that argument

[neon pasture]

but yeah, it doesn't really matter

[vale current]

And it makes sense

[white bough]

We could go to the other extrem and say that they are all classes(objects) as even a function is just a class with a __call__.

[neon pasture]

a function is not a class

[orchid thistle]

depends on if you mean "function" or function, imo

[neon pasture]

a function is an instance of the function class

[white bough]

Wait range is a type? Hm never noticed, sorry for that confusion

[orchid thistle]

there is a particular type named function and its properties are 100% objective
there is also the english word "function" which is a property of us and not of the language

[vale current]

FunctionType, yeah?

[orchid thistle]

yep ```py

import types
types.FunctionType
<class 'function'>```

[vale current]

Yes, but if you have a callable object, couldn't you consider it a function

[waxen fox]

i wouldn't really describe range as a function; more so an iterator

but to a beginner the word iterator is confusing

[vale current]

Isn't range not an iterator

It's iterable

[orchid thistle]

a range object is iterable, but not an iterator ```py

next(range(5))
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
TypeError: 'range' object is not an iterator```

[vale current]

You'd need to call iter on it to get an iterator

[waxen fox]

oh

til then

[neon pasture]

not really, no. Functions in python do a couple more things than just be callable (e.g. they have name, a code, any number of user-defined attributes, ...)

[white bough]

Just wanted to say that a function is nothing really special, and can have attributes set like an instance of an custom class and all the other stuff.

The main difference seems to be that the class of any function is builtin and writting def …, just creates an instance of it.

Which makes me wonder, what happens if you set an attribute via

def something():
    …

something.__class__.random_attribute = 7

[waxen fox]

imo i would say that functions are callable but not every callable object is a function, i'm probably wrong on that tho

[vale current]

Hm, I see

Although you can use a generic callable wherever you'd typically see a function, right?

[orchid thistle]

>>> something.__class__.random_attribute = 7
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: can't set attributes of built-in/extension type 'function'```

[white bough]

Its confusing semantics and in the end, a function is just an instance of an special class, thats what I wanted to say, everythings an object

[green warren]

Everything in Python is an object, including functions and types. Honestly this might be confusing to someone learning OO.

[neon pasture]

99% of the time, yes

[white bough]

Oh I should have guessed that, thank you though!

[neon pasture]

the most common example is a bound method, which is a different, ale quite cool, object

[white bough]

But going back to the base question, ala confusion, I actually thought range returns you an iterator, so what something actually returns can be confusing too

[waxen fox]

it would just be something.random_attributes iirc

[neon pasture]

yeah, range is a bit odd, since it's a special object which is not a generator, not an iterator, not a list, ...

[vale current]

Immutable sequence 😩

[neon pasture]

ye, it is closest to a tuple

[white bough]

Ok didn‘t think something basic and used everyday could still confuse me, but nice suprise

Wait is enumerate() actually behaving the same way as range?

[orchid thistle]

enumerate objects are iterators

[neon pasture]

honestly, pretending that range is not a sequence and just an opaque magic iterable is probably the best you can do.

[white bough]

Ok, then I am not lost completely, thank you

[neon pasture]

not like you are going to put a range object into much of anything other than a for loop for quite some time.

[white bough]

Slice?

[orchid thistle]

...?

[white bough]

This may exceed the channels goal, but as I am now curious any reason for the weirdness of range?

[orchid thistle]

well suppose you have range(1_000_000_000)
how do you propose representing that as a non-special type in a way that makes sense?

[vale current]

You can use a range more than once, right?

And they're immutable

[neon pasture]

neither of those things really matter IMO.

[white bough]

Was an answer to what you would put a a range object other than a for loop. But I am now not even sure if this is possible.

That revelation kind of made me question everything. 😅

Going back to the actual bot made question:

Something that just recently caused me problems again, actually because I wanted to explain it to someone and forgot the „weirdness“ is:

Class attributes and non-slotted classes

You can get the value of the class attribute from an instance by doing inst.the_class_attribute, but if you want to change it by doing

inst.the_class_attribute= 14

You will actually create a new instance attribute with the name the_class_attribute and not change the class attribute.

You actually have to set it on the class.

The confusing stuff can come from the way getting and setting seem to handle completely different.

They actually do not really as getting the class attribute first looks up the name on the instance and if it does not exist it goes to the class.

This can be used for nice things as a special way for default attributes, but can also be a huge footgun.

[violet prawn]

Do you guys offer help for assignments? I'm boutta fail at this rate ;oooo

[stone juniper]

this channel is for discussion about how to teach computer science, but you can open your own help channel from #❓｜how-to-get-help
rule 8 is relevant though

[desert patio]

if yes then i have no simple way of distinguishing functions from non-function callables

[vale current]

perhaps ```py
from types import FunctionType

def foo:
...

isinstance(foo, FunctionType)

[desert patio]

disadvantages of saying anything that is callable is a function:

1. beginners will say everything callable is a function

class A:
  ...

A()
```"it's a function"
2. arguments that don't even have a point
1. ```py
raise ValueError(7)
```"it's a function"

well according to english terms those are functions

but you still have to separate their terms in programming

[ripe folio]

This fails for methods

[vale current]

Hmm

[languid yacht]

What should a student do if he is in a big coding momentum but also need to do other homework.
I kinda want to solve this without losing momentum

[vale current]

!e ```py
class Foo:
def bar(self): ...
f = Foo()
print(type(f.bar))

[worldly dew]

@vale current :white_check_mark: Your eval job has completed with return code 0.

<class 'method'>

[vale current]

Ah welp

Well, there's types.MethodType, isn't there

[ripe folio]

Yes

[languid yacht]

if method or is a function

[junior whale]

Anyone know where i can find info on usage of this type of syntax in python? Im only asking here because i cant figure out how to search for it specifically.

message = client.messages \
                .create(body="Internet Gas Gauge in AZ report delivered to email",
                        from_='+1520 INSERT TWILIO FROM NUMBER',
                        to='+1 INSERT SMS DESTINATION NUMBER')

what i mean is the \ then new line .method call. I am a student and have never seen this used before.

[pure fog]

this syntax element is called "line continuation", you are continuing the line onto the next line

that said, this channel is about teaching python, not learning it. but maybe we can use your post as a starting point for discussion on arcane syntax elements and documentation/teaching thereof 😉

note that you can also wrap the expression in parentheses, then you don't need the \:

message = (
    client.messages
    .create(
        body="Internet Gas Gauge in AZ report delivered to email",
        from_='+1520 INSERT TWILIO FROM NUMBER',
        to='+1 INSERT SMS DESTINATION NUMBER'
    )
)

[junior whale]

Oops! Sorry for that then, but thank you for your reply! I appreciate it!

[hazy delta]

this discord is an amazing bit of programming pedagogy. just the vast number of people is enough to make it great. i'm really happy to have found it and thankful that people made it.

[misty dirge]

If your message was deleted, please keep in mind that this channel is for the discussion of teaching. Please keep it on topic.

.topic

[fleet apex]

**What makes a project a good learning experience for beginners?**

Suggest more topics here!

[ripe folio]

Ooh! That's such a good topic

At the end of it all, I think a good project is one that keeps the beginner motivated - that's all!

It should have a scope that meets the beginners expectations, and be at a level that they are kept motivated working on it.

If the project is too hard they will loose motivation, and like-wise if it doesn't challenge them.

[wispy comet]

I really like your answer!

[mossy belfry]

I feel like the real question is "how do you find a project for a beginner". lots of people in the "I went through a tutorial; now what" phase

[native hearth]

Slap them with a !projects

Lots of people dont even want to try those out either, at that point im not sure what else to suggest

[ripe folio]

Yeah but what makes !projects good projects for beginners?

[native hearth]

Most of them are fairly simple and possibly actually useful
Examples like madlibs, blog, game of life, etc

Theres also a bunch of links to other repos suggesting projects like build your own x

If not projects like these, what would you suggest

[mossy belfry]

!projects

[worldly dew]

Kindling Projects

The Kindling projects page on Ned Batchelder's website contains a list of projects and ideas programmers can tackle to build their skills and knowledge.

[mossy belfry]

I generally don't like long, barely-curated lists of stuff. especially lists of lists

the variance in difficulty of these projects is pretty high

[native hearth]

Maybe its time for pydis to curate a list itself and add it to the site

[orchid thistle]

i feel like something beginners don't tend to properly get is that errors are actually meaningful
like, it's not just a random set of rules, you get an error if you told python to do something meaningless
so they get an error and ask "how do i fix this", but the answer will always be "you asked for something that doesn't make sense and have to reconsider what you're actually asking python to do", not "oh just do this and this and the error will magically be fixed because python just sometimes decides it doesn't like you"

[silent goblet]

sometimes I think technology just doesn't like me

[vernal sentinel]

So you're saying that restarting my computer won't solve any code errors? 🤯

[gentle oriole]

Yeah, things like "How do I fix this error" always catch my eye. Makes me want to say "The error looks perfect to me, it's told you your code is wrong"

[orchid thistle]

oh that's a good response, i might use that

[heavy trench]

Literally so goddamn true, I work as a teacher on the side and sometimes my students frustrate me sooooo much bc they hinder themselves so much, its like they'll be building a complicated lil algorithm that actually unironically impresses me and makes me think they are getting the hang of this so well, and then they hit ONE error and instantly come to me w/o even reading or trying to decipher the error on their own, like half the time literally all i do is ask them to read the error and tell me what it means and then its always like "OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOH wait nvm OmegaLul", like you literally know how to do this homie just put some effor into it, I aint always finna be there for your ass

[iron cobalt]

what could be a good project idea but nothing 2 complicated still learning

[heavy trench]

What do you know already

[iron cobalt]

well a bit of apis i just finished a dc bot

ik a bit i think

but id also like it to be a bit of a challenge so i learn

[heavy trench]

You could create/use a database and create a lil rest api to interface w it mayb create some website w it

[runic sigil]

I'm pretty sure I (and probably you) have used this...if slightly modified!

[wooden grove]

!mute 436089796318724106 "6 hours" You've used up your mod attention budget for the day.

[worldly dew]

:incoming_envelope: :ok_hand: applied mute to @pearl orbit until <t:1651376733:f> (5 hours and 59 minutes).

[wooden grove]

Sorry @winter spear didn't mean to delete your comments as well.

[winter spear]

U r good

[runic sigil]

We get mod attention budgets? :P

[wooden grove]

Spend yours wisely 😄

[runic sigil]

Can I get advanced notice when it's running low?

:)

[fallow terrace]

that is extremely vague. you're pretty much always learning, and "complicated" is very different for different people.

[iron cobalt]

yea my comunication skills are as good as my python skills so u see bad bad

[fallow terrace]

but if you meant it that way, one of my go-to suggestions for beginners is creating Conway's Game of Life (whether that's in the terminal or a GUI or whatever)

There are quite a few different ways that you could implement it (which is just like anything else tbf), which I find quite nice

[runic sigil]

Sharks and Fishes is my go-to - a very similar idea.

(as in - it's a cellular automaton)

[misty dirge]

@iron cobalt @amber root @hexed sun please keep your comments relevant to the channel topic

[jolly night]

I've had that page there for seven years, and have gotten very few suggestions for additions or improvements. 😦

[odd mirage]

Hey there. I'm currently looking for ideas to design a 3-6 hours long beginner to intermediate python task so that students can work with e.g. numpy, pandas and maybe "intersect" some data to look for corellations.

I'm currently looking for datasets that would suit my need. I thought about using large bird occurences and UFO sightings and have a dataset for the latter. However, I have a hard time being creative to find specific tasks the students can look for.

Is anyone interested in giving me advice or trying to think through some possible tasks in that context? The datasets don't have to be final, I'm currently skimming kaggle to see what I can find.

[misty dirge]

@odd mirage numpy and pandas have a lot of functionality, so you might first decide what things you want them to be able to do, and work backwards to a dataset that would facilitate those operations

[misty dirge]

@empty niche I removed your message because it was off topic. See #❓｜how-to-get-help

[clear hazel]

I'm mentoring someone in programming but I feel like it's been tough going. They've been wanting to bite off some intermediate/advanced projects without knowing a lot of the basics. I recommend more basic resources to fill in the gaps but the mentee seems to not take a lot of interest. So lots of wheel spinning. What to do in this situation?

Ideally I suppose I'd find some intro to programming that is both fun/engaging and thorough with the concepts. But I have no sense of the former tbh -- I guess "kids these days" like YouTube videos but I tend to avoid them when searching for resources

When I was learning to program, the "boring" bits didn't seem boring to me (e.g. I remember thinking pointers were mysterious, but cool). So it's hard for me to get in the same frame of mind as someone who doesn't see it that way

[ripe folio]

What kind of resources and stuff have you had them do so far?

Courses and written guides? Leetcode/hackerrank and other problems? Walk-through tutorials? Online talks?

[clear hazel]

I've sent them https://automatetheboringstuff.com/ for some things

I think they're still struggling with some of the basics like variables, basic types, control flow so it's tough to move ahead to anything else

Perhaps they need some intro resource that is more project-oriented. "Here's how to build something cool (a game, say) and along the way you'll learn everything you need"

[ripe folio]

Yes-yes.. yes! I love projects. I think that they're singlehandedly the most motivating way to learn. The goal is to find something that they can reasonably complete with their current knowledge (I am including "things they're struggling on" as current knowledge) and have them go through it. Afterwards, it should hopefully be such an open-ended project that they can always improve it. Since the project is built on knowledge they knew about, they should be able to keep working on it past its original goal (or put the goal at a subset of the finished project).

Everything they learn should contribute to the project. After every new concept they should be left with a motivated feeling of "I can apply this to my project" - the same way an artist feels like they can use new techniques on their new/working paintings - so that they are more excited to learn.

What the project is about is not very important. Try to match it with other interests or something that they are curious about. Depending on age, you may want to prioritize making it something they can show off: someone a bit younger usually appreciates showing of their work to someone they care about, while someone more mature and older can take pride in having made what they did (without needing to show it to someone else). This means that for someone younger you'll especially want something pretty interactive, while for someone older it is fine to create a program that only makes sense to other programmers (Python modules, web APIs, general computations, etc.)

I think finding something interactive is beneficial for everyone, but more-so a younger learner.

A nice typical project could be a username generator or password generator. That will have some loops and different types. As you go along they learn about making the program better with classes, adding save-ability as the learn about I/O, and making it prettier if you teach them about TUIs.

[clear hazel]

Yeah I think the challenge is to make it simple to make a super-barebones version, and then add more functionality as you learn more

Like a text adventure game could start with just input() and print() and expand from there

[ripe folio]

I like that, have them start it as moving a character on grid in their console.

It'll bring them up to lists and dictionaries to store information. Later you can convert this to classes and show a good example of why they're necessary.

The game can easily scale up to become more advanced. Another idea is having them learn about text processing by adding levels loaded from text files.

[clear hazel]

That seems like a cool idea. If you or anyone knows of any existing resources that do something like this (and by "like this" I'm casting a very big net around all beginner projects), I'd be grateful. In this case it's not worth it for me to make a custom curriculum

It's more informal, like I want to be able to recommend them helpful resources but it's just for fun in my own time, not something I'm paid to do/in any official capacity

There are so many low-quality tutorial sites etc out there, so I feel like one thing I can do as a more experienced developer is help separate out the wheat from the chaff

[desert patio]

also adding to this:

b = 7
a = b # is not the copy
a += b # is the "copy" (a + b)

[pure fog]

if it makes you feel better, this is probably a good learning experience in and of itself

i think a lot of young people who write code fall into a category of "intelligent and eager/impatient" - they need to learn how to think with intention, and to rely less on their natural intelligence to carry them through problem solving

[clear hazel]

It is like Kahneman's System 1 and System 2 idea

[pure fog]

maybe, yeah

rather than seeing "error" and instinctively reacting "idk", you have to force yourself to read the error

it might also be a kind of psychological avoidance mechanism - you don't want to examine your own apparent mistake

[clear hazel]

Yeah that's a common theme of debugging too

[pure fog]

indeed, and reading docs: don't skim, read carefully

[clear hazel]

You thought it would work, and now it doesn't, so you were wrong somehow... so somehow you have to go back through all of your steps and question every assumption

[pure fog]

i think the idea that "you have made implicit assumptions which need to be questioned" is somewhat advanced

learning to identify and quickly narrow down those assumptions is definitely a skill that comes with practice and experience (and advice/help from others)

but first you need to just read the damn text and stop to think about it for a second!

[clear hazel]

Yeah your brain is definitely the best debugging tool

Sometimes I see beginners do something like:

# assume a is defined
b = a.foo()
c = b.bar() # this line errors, maybe a type error

And once they insert a print(b) it all clicks into place

They assumed it had one value but it has a different one, so they need to look at foo() and not bar()

And then the next logical question is "okay so what code do I write next??"

[pure fog]

it's probably important to have some kind of mental model of "expression evaluation"

you basically have to be able to follow the evaluation graph backwards from the location of the error until either you find the problem or you reach the end of your code (and then need to backtrack to another branch of the evaluation graph)

[clear hazel]

yeah I think the backwards bit is what's so challenging

Your code runs forwards but you have to piece together what went wrong backwards...

[noble aurora]

a thing that really works is asking them to "act like the computer"

[spiral warren]

can anyone help me with my exam

i have a ex pylab code that i have to eksplane

[upper flint]

We don't provide help on exams

!rule 8

[worldly dew]

Rules

8. Do not help with ongoing exams. When helping with homework, help people learn how to do the assignment without doing it for them.

[spiral warren]

ok sry

[steady raft]

Hello, I usually work with R/Rstudio and I struggle to switch to Python. I mostly do data analysis. Does anyone knows a good transition method ?

[weak mortar]

Hi, I'm trying to create a fun quiz for my students who have 0 experience of coding, so any suggestions? :)

[spare jackal]

This is great book https://wesmckinney.com/book/

[misty dirge]

What is the quiz intended to be about?

[weak mortar]

Could be the basics of Python. Any topic

[misty dirge]

you can ask them to write a function that takes a list of ints and returns the maximum element using a loop, and ask them to spot syntax errors in some basic code.

[weak mortar]

Good idea actually, thanks

[misty dirge]

@analog glen be sure to DM @viscid mural if you are concerned about a scam

[mossy belfry]

I think a lot of modern computing is filled with inscrutable (or worse, misleading) errors. it's also really unclear what the traceback is for when you first start. so most people, especially the tech savvy ones, have learned to ignore the contents of errors

I think a common thing when I'm helping someone less tech savvy (my parents) is I found out they went down some pointless rabbit hole for 2 hours because they were misled by an error message

[jolly night]

i don't think tech savvy people ignore the contents of errors. I think it's that they've learned to identify the likely useful information in the error.

[misty dirge]

I agree. I usually call this the "salient part" of the message.

[steady raft]

Thank you I'll give it a try 🙂

[runic sigil]

I'm not sure this truly belongs here; I'm sure it can be leveraged for pedagogical purposes, though:

(Cunningham's Law)

[desert patio]

i like this

[runic sigil]

I reckon it can be used as simply as asking if the questioner meant something utterly ridiculous, and letting them answer their own question in the process. (I'm not sure I don't already use a mild form of this sometimes }:) )

[ripe folio]

I think there is a discussion to be had about how to write good error messages ☝️

[desert patio]

i wanna know how to do that

[weak mortar]

Will you help? Help me write a function that will reverse the sentence. Example: Hello world people. Very good. Conclusion: people world hello. Good very.

[jolly night]

I think this is similar to how people often think docs are bad: it comes down to guessing what people already know, and what they need to know. Those guesses are often wrong.

[native hearth]

you could write error messages that solve the problem themselves, some people still wouldnt read them

how many people in pygen post error messages that explain exactly what the issue is and sometimes even how to solve it, asking for help

[ripe folio]

Yeah exactly

Do you all think error messages should be technical?

[runic sigil]

The problem with non-technical error messages is that they aren't specific. The problem with technical error messages is that general user doesn't always know the distinctions made by the technical language and often bewilders rather than helps. I wonder if there is a usecase for having "simplified" or "technical" versions of error messages, selectable by some setting!

[pine quartz]

Potentially the problem then is that users hitting the simplified errors will report them anyway, but no longer include the detail needed to solve the issue.

[cerulean bough]

Maybe we shouldn't rely on users to report the issues.
When some error occurs, a message is displayed and a log is created.

[orchid thistle]

that's pretty much how Rain World modding works except we don't have anything to display a message so if an error occurs the game will just stop working
but it generates a lot of rather technical information about what's going on into log files

i think for RW it's a pretty reasonable approach because nobody ever has any non-technical information about what went wrong that could reasonably be presented
although if random users try to read the logs they generally end up extracting worse information than if they had completely guessed

that's the general problem: information is useful if you know what you're doing, but detrimental if you don't know what you're doing and don't know that

[cerulean bough]

I think more reasonable would be to give some feedback to the user :D.

Even "the game stopped working" is better than nothing.

[orchid thistle]

that's not even intentional, it's just that once an error happens the game is usually in an invalid state now and will keep trying to update itself and not getting anywhere so nothing ends up happening

[cerulean bough]

That seems like a bug :D.

[orchid thistle]

no it's just how rain world is written
there is no error handling anywhere, other than Unity catching exceptions which doesn't actually stop the game

[cerulean bough]

"no error handling anywhere" looks like a bug to me :D.

[orchid thistle]

i think this is the main problem with errors
to deal with non-trivial errors you need technical information, but if you can't deal with technical information then you'll get confused

[orchid thistle]

well if there are no errors then you don't need any error handling

[cerulean bough]

once an error happens the game is usually in an invalid state
pretty hard to not have errors

[orchid thistle]

well yes, once an error happens more errors will happen
but if you just never encounter an error in the first place, then the game runs fine

[cerulean bough]

"dress for the slide, not the ride" as they say

[runic sigil]

This is why users get confused - domain-specific terminology! I was trying to work out why you'd dress differently to go down a slide than to get to the park to go on the slide ;P (I assume it's from dirtbiking or some such?)

[cerulean bough]

Yes, it's a saying among motorcycle riders :D.
And that's why we should have different messajes depending on the audience: normal users get something generic so they know it's something they can do better (or not), and we log the tech details so the devs can fix the issue.

[covert prawn]

hi i had a question is there any way to compile pyhton code to an executable like c++ ?

[formal flower]

Yeah! U can use pyinstaller

or some like that, But I don't find it useful

[covert prawn]

i found an alternative called nutika (i was wondering if such applications are actually useful in real life applications)

[stone juniper]

this channel is for discussion about how to teach computer science, but you can open your own help channel from #❓｜how-to-get-help

[somber rain]

Anyone know the specific legals behind showing a Udemy course to a small class of students? I'm just wondering if my teacher can do so without getting into trouble since I couldn't find anything mentioning this online...

[misty dirge]

is there a reason you care if your teacher doing this is legal or not?

[somber rain]

Not really

I'm just wondering

He hasn't actually done it; but could he do it?

[misty dirge]

As a student, when you enroll in a course or other content, whether it’s free or paid content, you are getting a license from Udemy to view the content via the Udemy platform and Services, and Udemy is the licensor of record. Content is licensed, and not sold, to you. This license does not give you any right to resell the content in any manner (including by sharing account information with a purchaser or illegally downloading the content and sharing it on torrent sites).
showing the content in another course could be construed as reselling the content, I guess.

In legal, more complete terms, Udemy grants you (as a student) a limited, non-exclusive, non-transferable license to access and view the content for which you have paid all required fees, solely for your personal, non-commercial, educational purposes through the Services, in accordance with these Terms and any conditions or restrictions associated with the particular content or feature of our Services. All other uses are expressly prohibited. You may not reproduce, redistribute, transmit, assign, sell, broadcast, rent, share, lend, modify, adapt, edit, create derivative works of, sublicense, or otherwise transfer or use any content unless we give you explicit permission to do so in a written agreement signed by a Udemy authorized representative. This also applies to content you can access via any of our APIs.

that probably prohibits it as well.

I am now Stelercus, Esq.

[weak mortar]

I am now Stelercus, Esq. what does this mean?

@somber rain why?

[misty dirge]

"Esquire" is a title that lawyers sometimes put after their names.

[weak mortar]

:D

[somber rain]

Alright, thank you!

[weak mortar]

Can I ask why?

[somber rain]

Just generally wondering

[misty dirge]

because I read a terms of service and gave my uninformed opinion.

[somber rain]

Like, literally no specifics

[weak mortar]

no, I was asking happyalt, why the want to know

[misty dirge]

I already asked that further up

[weak mortar]

your sign was fine ;p

I know

[weak mortar]

I don't buy it :)

but ok

[somber rain]

Well, what do you expect?

[weak mortar]

I meant to say nvm right after that, but I was timedout :)

[misty dirge]

@somber rain if you're trying to start a udemy resale pyramid scheme, you have to pay the two of us off now to keep us quiet.

[somber rain]

Lmao no no

[weak mortar]

:D

HappyAlt, I expect you can get in trouble if you sell existing lessons for money

if you do it for free, I'd think there wouldn't be a problem

[somber rain]

I'm not planning on selling any lessons

I was just thinking of gifting my teacher a udemy course so he can (possibly, if legal (which I now know is not)) show it to the rest of the class too

[weak mortar]

that sounds great

he doesn't need to show the video to pass it on

he can learn from the course and pass it on

[somber rain]

Yup

it's a shame though... the animations in the course very very helpful in explaining what was going on

[weak mortar]

dude/dudette you actually worrying about gifting a professor a course and them getting in trouble if they share it?

respect

but don't worry, it's not the end :)

[somber rain]

ty, but pretty sure I'm just covering my own butt too :)

[weak mortar]

unless you are from usa and you or the professor copies everything. you should be fine :)

[somber rain]

1. Not from the USA
2. Professor will definitely copy everything

Lmao this is high-school, not even Uni or something

[weak mortar]

I'd say you can gift your professor

what they do with that - is on them

[somber rain]

The only reason I want to gift them a python course in the first place is because the guy makes us do everything in VB.NET

[weak mortar]

I feel you

[somber rain]

And I don't wanna make a Textbox to get input and a label to display output just to do:

a = 7
print(a)

[weak mortar]

I don't agree with you here necessarily :)

[somber rain]

How so?

[weak mortar]

why don't you want to do this this way?

[somber rain]

Because we've made the exact same 2 or 3 programs for the past three years

[weak mortar]

ah, ok :)

yeah

scratch my assumptions

[somber rain]

I've literally just taken my past year's submissions and submitted them to him again, just to get full marks

(but shhh on that)

[weak mortar]

I don't think you've been honest about yout situatio/position :)

[somber rain]

Well, to be frank, and with all due respect, I asked him a year ago why didn't just make a function out of something (can't remember what), and he said he "didn't wanna lose people". So, he spent the next 2 minutes copy-pasting the same lines of code and changing up a character here and there to make it work

And I was wondering: if he's afraid to lose people because functions might look scary or something VB (or whatever else), why doesn't he do them in a language most people consider to be more syntactically pleasing... like Python

[weak mortar]

:)

https://www.youtube.com/watch?v=ZLyP_SeQ9p8

sorry, I can't help you with your situation

I also don't think you've been completely honest

anyway goodluck :)

[somber rain]

That's alright... Although, I'm confused by what you mean that I haven't been completely honest... Do you think I'm going to re-sell the course or something?

[weak mortar]

nah, you were saying you like your professor and your wanna buy him something

it's clear you don't think much off him

[noble aurora]

maybe you guys should move to an off topic channel

[somber rain]

Good idea

[weak mortar]

yep

[somber rain]

@weak mortar #ot1-perplexing-regexing

[sage jasper]

@whole nebula hey this is the wrong channel, feel free to ask in #python-discussion or #❓｜how-to-get-help

[sturdy ice]

Is it safe to say that it will generally take time for a person to practically solve Hackerrank/Leetcode problems even if they have a CS degree? Because you generally don't learn to approach these sort of problems in college. I am getting back to programming after about 3 and a half years.

Is there a specific learning method which focuses more on problem solving than SE done in the industry?

[misty dirge]

if you've taken an algos/data structs course, you'll probably be able to better understand the question in formal terms. but I'm not really sure how one gets better at answering them in general.

[sturdy ice]

I took both Data Stuctures and Algorithm Analysis/Design

That too from Pakistan's only good theoretical comp scientist, well he is the only and taught a lot of CS students

But I am not going to say I am good at Algos, I feel rusty

I can't figure out a learning path to be precise

My current learning path is

• CS61a from Berkeley to refresh Python quickly
• Python training at work for 2 months, I requested to move to technical side from management after 2 years

[orchid thistle]

there's always the fully general approaches: do the thing a lot, look at things made by people who are better at doing it than you

also break down what exactly the thing is into smaller things that you can look at individually

[sturdy ice]

This is my focus and objective right now. You can say I have some PTSD from working at a startup, and it left me scared and paralysed to even attempt breaking down a programming problem

But yeah makes sense

I just kind of need validation as to if I am on the right learning path

[orchid thistle]

...i meant that as an approach for getting better at things rather than advice on problem solving but now that i think about it it does actually work on both of them

[sturdy ice]

Yeah it does, Feynman made it popular and now Elon has made it mainstream. The First Principles method

You keep breaking down a problem to the most fundamental facts which you know hold true

And then you build up your solution from the very bottom

Excellent way of learning

[orchid thistle]

...fundamental facts? huh
what i was going for was either of:

1. generally breaking down a task into smaller tasks, that aren't necessarily fundamental but that are easier to address than the entire task
2. taking a programming problem and breaking it down into primitive operations

i guess breaking things down into parts really is a very generally useful concept

[sturdy ice]

Hmm yeah, I have been doing that, but I feel that I have to ask someone if I am right rather than watching the code run successfully. To know that my solution is elegant and simple.

It's just that I had a very bad experience when I got my first job and started working in React. There was no leadership or mentors. React also didn't make it easy.

The learning was without any goal or objective in mind. I didn't people to talk to either. I really am enjoying my time here because people actually care about each other.

[orchid thistle]

simplicity/elegance is a bit difficult to check
i guess one thing you could try looking at is, how similar is the code to the initial idea you had of what you want to do

if you have a problem that's just "do this, that, and then this other thing" and you can translate that directly into code, then that's probably a pretty good way of doing it

although i guess in that case it's less of a "problem" and more just a specification

if you have something where it's properly difficult to work out what you're even meant to do ("here's a graph with a million nodes, check if there's an eulerian cycle") then it is pretty difficult to tell whether there's a better way than what you've come up with

[sturdy ice]

Or how different it is as you go along and learn new, more efficient ways of solving

[sturdy ice]

Yeah my heart wants to solve these kinds of problems as a fantasy but my mind wants me to focus on the fundamentals

This all makes sense

I am glad I asked my question here, I was a bit nervous

[orchid thistle]

oh i just remembered another approach that sometimes works
if you want to write a program to solve a problem, think about how you would do it manually on paper

it's not a perfect analogy because computers are way way bigger but ultimately the idea of imperative computation is the same whether it's a human following a series of steps or a computer following a series of steps

[sturdy ice]

Yeah makes sense, okay I will always keep paper right next to me

Thanks @orchid thistle

[orchid thistle]

you don't need actual paper, just the idea of paper

although i guess you could solve an actual instance of the problem by hand if that helps with working out how you would do it

[sturdy ice]

I prefer paper, I like to feel and create not just do it in my head

[orchid thistle]

fair enough

[sturdy ice]

This will be my next goal after I learn to break down the basic problems because you can't just go and code a seriously hard problem without having an idea how to approach

[desert patio]

if you ever wanna search a python object in the API docs, add "docs" after your search if it isn't there already

without "docs"

with "docs"

[native hearth]

Pydocs SEO might be the biggest pain point when teaching someone python

Constantly tell them to look things up, do your own research, try things to see what they do, etc
But when they do try to google things, first result is that shit w3schools site

[runic sigil]

My usual trick is to add site:python.org. docs is definitely a shorter solution :)

[wide folio]

Guys can anyone help me understand Microservices so it sticks in my head. I need to memorise it tomorrow to explain in a job interview. Anyone have a good way of putting it ? And how it differed from macros/monolithic systems?

[fleet pulsar]

It's even worse in countries where english is not that widespread

Imagine having about 3 russian sites about w3schools-tier of bad pop up BEFORE w3schools BEFORE the actual docs

[weak mortar]

Why does w3schools suck?

[native hearth]

It has incomplete, out of date, sometimes just wrong info

[weak mortar]

Can you give me an example?

[stone juniper]

stolen from a staff thread for bashing w3

[misty dirge]

!e max(5, 10, 15)

[stone juniper]

point is that it's not an example of what the sentence right above says

[misty dirge]

true

anyway, the bad learning website can't hurt you here

[sage jasper]

is the SEO better for docs for other languages?

[native hearth]

Very quick (1 search lmao) google says yes

[sage jasper]

bc i agree, it's quite annoying that i have to go out of my way to find what i would normally expect to be result #1

[native hearth]

Even python docs arent that great to navigate tbh

Whats up with shit like this

[vale current]

Lmao

[native hearth]

Imagine being a beginner trying to find things by yourself

[vale current]

Maybe a site which filters and searches the docs for you would be cool

[native hearth]

No wonder they dont read docs

[runic sigil]

There is a readthedocs site that tried to replicate the Python docs, and be "a better source of documentation"...but I can't tell if it was abandoned, or is just poor ;)

[glass hamlet]

hey guys, im learning django and in a tutorial the guy said templates arent used anymore, but on the other tutorial, half of it is spent on templates

what do you think?

[runic sigil]

I think you're looking for #web-development :D

[glass hamlet]

sorry mate, my bad

[wide folio]

Guys can anyone help me with an interview question?

Why is assuring system security a more challenging problem than assuring system safety?

[misty dirge]

you might ask in #cybersecurity. evidently, they expect you to know a technical distinction between security and safety.

[pure fog]

not for C/C++ or Java

is there a ddg "bang" for python docs?

[sage jasper]

what's that

[fallow terrace]

yes, !python

[fallow terrace]

https://duckduckgo.com/bang

[pure fog]

that's great, i will definitely use it!

i'd love similar for other programming languages, especially those without good built-in search functionality in the docs like lua, ruby

(ruby does have a search function on the docs site, but sometimes i wonder if site:ruby.org is more useful)

[sage jasper]

is there maybe an equivalent easy way for non ddg users

i guess just appending "docs" like the user above suggested

[fallow terrace]

on the page I linked above, there's a whole search area where you can search any bangs fyi

[pure fog]

aha, very useful

[stone juniper]

...cant believe i never thought of this

[fallow terrace]

I'm surprised I didn't know bangs existed before 😳 1 more reason ddg is better than google

I wonder what could be done to improve the SEO of the Python docs.

[sage jasper]

if donations will make a difference i'll donate

[misty dirge]

I suspect the PSF would sooner put donations towards having full time cpython devs or keeping the pypi up

[sage jasper]

probably

[misty dirge]

Also at least one of the steering council members is a Google employee 🤔

[sage jasper]

is it actually a matter of $ though? or just knowing how to game SEO? (or are those 2 things the same)

and if off topic we can move

[fallow terrace]

I think it's the latter but I'm not sure. Perhaps it's simple things. web.dev is complaining that there is no <meta name="description"> on docs.python.org/3/, which should be a fairly trivial fix.

[analog sorrel]

Safari instead

[analog sorrel]

Psf is counting donations in pat times

Part*

[analog sorrel]

Ok

[fallow terrace]

I dunno, I don't use Safari

[vale current]

Open an issue about it perhaps

Or well, mailing list stuff

[weak mortar]

we need to open a therapeutic industry for programmers facing illogical errors.

[rain plume]

!rule 5
@brittle sand's

we don't do that here
refers to the server, not the #python-discussion channel

[worldly dew]

Rules

5. Do not provide or request help on projects that may break laws, breach terms of services, or are malicious or inappropriate.

[rain plume]

it's also off-topic to pedagogy, which is about teaching

[sage jasper]

@stiff halo please move off topic messages to off topic: #ot0-psvm’s-eternal-disapproval

[alpine hare]

Hey guys, not sure if this is the best place to ask this, but I couldn't see a better fit.
I've been working as a system / security engineer for the last 4 years, and I have constantly started beginner python courses. I have always went through the introduction to the language and some frameworks, but nothing big.

As my next step, I want to really UP my python skills and I was looking for courses that are not beginner oriented. I cannot stand going through another introduction to the language and I'd love to be able to apply what I know (not much tho, I am still learning)

So, if anybody has any courses that are fit for my needs, do share.
Thank you!

[cerulean bough]

@alpine hare this channel is for teaching Python, not learning.
If you think you are beyond basics then it's time to start a project.

[alpine hare]

Oh, excuse me, I really thought teaching and learning are connected 🙂

[cerulean bough]

They are, but you are at the other end :D.

Feel free to DM me if you need help with starting up, no strings attached.

[robust bane]

As sailowcat has said, your question is off-topic for this channel.
We have a curated list of resources at https://www.pythondiscord.com/resources/ .
With that said, if you feel like you're past the beginner stage, I suggest that you start a project. It will help re-enforce what you've learned as well as provide you with a fun and interesting challenge.

[fluid pebble]

whats a main routine, is it suppose to call your functions?

Im working on a flowchart and need help

[ripe folio]

This is not the topic of this channel, see #❓｜how-to-get-help

[fluid pebble]

ok 👍

[tiny swallow]

When we use calculators like mathway, symbolab, or wolfram alpha, I have always wondered how these programs are able to show things like: Steps taken ?

How do programs do this? I just think its so cool and I wanna try to do something similar for like an independent project.

[misty dirge]

a recent episode of TalkPython is partially about this: https://talkpython.fm/episodes/show/364/symbolic-math-with-python-using-sympy

[tiny swallow]

Awesome! Thanks--ill check it out

[median folio]

Thanks

[weak mortar]

Guys who can help me or check with me a code in C cuz im working on a school project i need to deliver by tuesday max 400 lines . Thanks

[proper mason]

This is a Python Server, try https://discord.gg/programming

[weak mortar]

Thank you so much mate

[rain plume]

alternatively, a c-specific server recommended by this one is https://discord.gg/gWXQJdm57f

[long heath]

What's wrong in the final min, max example?

[oak breach]

it says it takes an iterable

[pure heron]

min and max take either an iterable or 2 or more items to sort. The prose above the example explains the first way of using min and max, but the example shows the second way.

[woeful garden]

Hi. Are you guys able to answer questions on Kademlia?

[misty dirge]

This channel is for discussion the methods and practices of teaching. If kademlia is related to that, just ask; don't ask to ask.

@woeful garden kindly remove your comment and re post it in a channel that more closely relates to it. Probably #algos-and-data-structs

[valid marlin]

Here if we write just 2 not str(2) then why it's showing errors, why can't we just push numbers

[noble aurora]

you can, it's just erroring when you try to print it out because you can't concatenate an int to a string. btw, this is the wrong channel, see #❓｜how-to-get-help

[misty dirge]

@empty dragon your comment did not appear to be on-topic, so I have removed it. Please refer to the channel description.

[split cliff]

Hey guys, how are you? I've been learning some programming languages for over half an year, mostly python, but I kinda feel like I don't know the context and processes behind at all; I feel like I'm programming only for "knowing the most languages I can". Do you recommend me any course about computing in general, like a CS course (idk) or something related?

[misty dirge]

if you're "knowing the most languages I can", it sounds like you've been going for breadth over depth. the basics of any language are going to be easy to pick up, especially if you've learned other languages. but that doesn't mean you've mastered any of them, or more importantly, know how to accomplish anything marketable in them.

you might reframe what your goal is. If you're past the point that trying to learn "python" is getting you anywhere, you should probably try to learn a specific domain of programming, like web development or backend development.

[split cliff]

Since I would like to know more about development in general (how a whole app/web site is built) what do you think I should learn?

[misty dirge]

this channel is actually for talking about programming education itself, so you might head over to #web-development

[sacred violet]

If you guyse watch tutorials on YouTube, what do you like about them and dislike about them?

I'm also interested in learning about approaches in teaching to code for absolute beginners.

E.g. there are those robot games where you basically can control a robot but without all the overhead of a language. So you can focus on the basic "mindset" of coding before using a language. Is that a good idea?

[stone juniper]

I don't like videos because understanding code is really about going back and forth as needed while reading it carefully; it's virtually impossible to do that watching someone's video.
I do like those robot games for teaching procedural thinking to absolute beginners because

• the language they have to learn is tiny
• they get immediate visual feedback on what their code does
• they're manipulating data that is effortlessly intuitive

[misty dirge]

Videos are okay for getting a general sense for something, but with programming, you can't just passively absorb the knowledge, no matter how closely you pay attention to the video. Much like with math or a foreign natural language.

[sacred violet]

Yeah but the benefit is to have the code and an explanation. E.g. I don't see much value in a book. I once, years ago, tried an interactive thing which I guess is nice? Cant really decide.

Sure in the end it's just practicing it as always.

[noble aurora]

there's nothing preventing a book having code and an explanation

[sacred violet]

Ive probably read more than watched, so sure

[misty dirge]

Not sure I follow. Books can have code and explanations, and often do. There are entire publishing houses that just produce programming books.

[uneven tinsel]

for development in general, it'd be good to learn how to use git (https://git-scm.com/book/en/v2) or another SCM (but probably git 🙂 ), data structures and algorithms, testing (unit, integration, system testing), debugging techniques... probably a few other things I'm not thinking of

but that's a good start

[sacred violet]

Does anybody know a good "robot game"?

Do you think it's better to use something like Karel i.e. a robot game focused on the very basics and extra simple syntax vs. e.g. the same kind of game using something like python?

[native hearth]

not sure what Karel refers to, do you have a link?

depending on who you're teaching, gamification may not always work

i would definitely not appreciate someone trying to teach me, a 25 yo grown man, via silly games, especially for something as inconsequential as syntax
would definitely put me off

[sacred violet]

https://github.com/fredoverflow/karel It's about teaching coding basics for people who never coded before.

[noble aurora]

right, but i think this is targeted towards people that are more like 10 or younger

[pure heron]

There's a game called Human Resource Machine that has you programming progressively more complex programs in a limited assembly language

I, as a 35 year old grown man, enjoyed it 🙂

[stone juniper]

have you tried Exobots?

Exabots? something like that

[pure heron]

I haven't, no

It's good?

[stone juniper]

it's great

parallel assembly to network little robots to do tasks, with leaderboards

would not recommend for beginners though, because I'm not cruel

[pure heron]

Human Resource Machines starts out pretty beginner approachable, but the difficulty rapidly climbs. One of the later levels requires you to implement a sorting algorithm, I remember...

[sacred violet]

what? No, why do you think that? Even the children langauge scetch or what is called is used in the first week of standford. Karel is also used by several universities. I guess one doesn't spend half a year with it though but simply to get started.

[noble aurora]

you literally called it "the children language"

[sacred violet]

What? I called sketch a children language becuase it is focused on childrend, I didn't call Karel one.

[stone juniper]

looked it up- it's called Exapunks

i was close

[tepid edge]

i'm not familiar with copyrights rules and restrictions (there is no such rule in my country). i want to know for example if i want to create an open source translation of (https://plato.stanford.edu/contents.html) what should i do ? just write a sentence that the copyright is reserved for the main site ?

[cerulean bough]

I don't find the license. I think you don't need a licence to do derivative work, but if you want to be sure write them and ask.

[shell ravine]

@tepid edge I think you are not going to have a problem if you are not making money out of it.

[pure heron]

legally speaking, you need permission from the copyright holder of that page if you want to distribute your own translation of it

[languid yacht]

Is it impossible to learn good code by yourself?

Because I can't learn from my mistakes if no one points it out

[weak mortar]

I started learning python from yst n I’m a lil confused about variables and data bases and what brackets I should use in python

[silent walrus]

By looking at how others solved a problem you can see different ways of solving an issue, but also just asking people here if there are ways to do something better

[misty dirge]

@weak mortar this is a discussion channel. See #❓｜how-to-get-help, and remember to ask your actual question, rather than asking to ask.

[weak mortar]

You can, and you'll do

[misty dirge]

all learning is self-learning, in my opinion. and a lot of mistakes will cause your program to fail, in which case Python will tell you why it failed, at least at a high level. and you can ask for code reviews here.

[hardy beacon]

i feel like some people want to learn programming without doing any hard work. like they want to learn how to program and stuff, without having to do a bunch of trial and error. or perhaps they just dont realize that they cant be spoonfed all the info and know how to program. but regardless, how do you teach these people how to program?

[tardy plaza]

How do I get help with Java?

[stone juniper]

i'd ask on a java server

[stone juniper]

first, don't spoonfeed them
I like to ask them questions that they have to learn to ask themselves, like "What does the error message say?", "What is on the line that it's pointing to?", "What did you change?", etc.
If it's a question that can be answered with documentation, I'll answer it and then link the relevant place in the documentation.
I also try to make them figure out exactly what question they're asking, because understanding the problem is always the first step in fixing something.
Basically: still help them but in a way that teaches them how they can help themselves

[hardy beacon]

hm thanks, that makes sense. but what if they have about no programming knowledge. how do you start people off in programming?

[kind flint]

Hello

I am new to python

[stone juniper]

I'm less familiar with that since most of the people I help are on this server and so probably have at least a little experience by the time they find their way here. I think structured tutorials are good. I like DSLs like those little robot games that teach people about procedural thinking. See up here for more of my thoughts on that ^

[hardy beacon]

i see. thanks for the help!

[cerulean bough]

We explain the fundamental concepts, with examples, that's not spoon feeding.
We still ask questions when the subject relates with knowledge they may already have or for feedback. This helps with keeping them involved.

I guess first you need to see if they want to learn, because some don't, and in that case I don't think there's something we can do.

[elder mauve]

Hello! I have a doubt. These last weeks I have been doing Python exercises in HackerRank and I have learned a lot. I checked for similar pages and I got a lot of results. I don´t want to check page by page, so I want to ask if you guys know a good page that is similar to HackerRank?

[pure fog]

tell them that there is no alternative to working hard, and that shortcuts now will only slow them down later. if they ignore your advice, then there's nothing you can do for them and you shouldn't waste your time helping them. they'll eventually have to learn the hard way... can't save everyone from themselves all the time. some people have to figure it out on their own.

that's true for a lot of things in life, not just "school"-related things

other times, you might need to present things differently. ask them "why don't you want to work hard? how do you expect to learn otherwise?" -- that kind of thing. sometimes people just need to hear it stated a specific way before it sinks in

[misty dirge]

@pure fog I think "working hard" can have a negative connotation that doesn't necessarily apply here. Learning to program does require a lot of work, but it doesn't have to be "hard" as in "agonizing".

[pure fog]

that's true. "requires concentrated mental effort and deliberate practice" is maybe a better expression of the idea

[misty dirge]

I will now require you to say that whole thing every time, and I'll act confused if you say "hard work".

[native hearth]

but what about learning x the hard way

[misty dirge]

if two procedures for learning something have the same educational outcome, but one is more difficult than the other, the less difficult one is better.

[native hearth]

i wonder if thats why the server doesnt recommend "Learn Python the hard way"

[misty dirge]

no, the problem is that the information in that book is poorly presented, and the author is condescending. and has an irrational hatred of python 3.

though with beginners, I can appreciate that it's difficult to make anything that feels worthwhile, and it's important that there's some kind of satisfaction loop.

[pure fog]

what i described is the hard way -- the intended purpose of those books (even if they fail to achieve it) is that they require you to spend time with the code and apply your critical thinking skills from the start, rather than just delivering information to you

[weak mortar]

When learning js is it a good idea to get a good grasp of document object model

[native hearth]

Absolutely, but that is not on topic for the channel, you can ask for more details in one of the offtopics #ot2-never-nester’s-nightmare

[hardy beacon]

@cerulean bough @pure fog thanks for the help, will keep it in mind

[dull basin]

does anyone know a good website that can give me project suggestions to work on?

[upper flint]

!kindling

[worldly dew]

Kindling Projects

The Kindling projects page on Ned Batchelder's website contains a list of projects and ideas programmers can tackle to build their skills and knowledge.

[spark flame]

Anyone know what's this?

logics

[misty dirge]

@spark flame @green belfry your messages are off-topic for this channel. if you want to discuss boolean logic, that would be relevant in #algos-and-data-structs

[green belfry]

Oh alright

I thought this channel is for kind of thing

[misty dirge]

it's for the discussion of teaching itself

so mostly how to teach OOP and type theory

[cosmic hill]

how do people keep their heads on top of object-oriented frameworks? Seems there's constantly more things I don't know or am missing to do what I want

Using django for my dissertation project

[misty dirge]

as opposed to functional frameworks? what other languages do you have experience with?

[cosmic hill]

Front end mostly, and PHP

[misty dirge]

ah yes. internet perl. try going to #software-architecture and give an example of something you're finding confusing in django as it relates to OOP.

[dark lintel]

Hi, I learned the basics of python (and a little bit of objects programmation as well). I did a coursera class plus some courses at my uni. Now I don’t know how to continue, I want to improve my coding skills and also learn machine learning and ai but I have no idea on how to proceed, I’m a little bit overwhelmed. Do you have any advice for my situation ? 🙂

[somber venture]

Try doing projects @dark lintel

[dark lintel]

Ok thanks ! Do you mean projects on my own or with other people ?

[thorny mica]

Do projects on your own with gradual difficulty/complexity increase

[weak mortar]

How many hours should you code as a beginner

What is the minimum

[noble aurora]

0

[lunar wave]

Studies and averages would suggest that the average attention span is productive at learning through applied study for about 15-25 minutes in one continuous go. This, of course, varies for each individual.

So how many hours should you code as a beginner? You should set aside an uninterrupted block of time, apply yourself to learning a concept in code, and stop the moment you feel like you aren't making more progress (or not picking up what you are reading/writing/listening to). Take a 10 minute break. Start again.

The total time you spend doing this is up to you.

That's one theory, anyway.

[buoyant zealot]

since python allows one as programmer to get away with a lot of stuff that could be considered dubious in any other language, like no type declaration and many other things... what would be the best programming language to learn to get rid of some of those bad habits that could have manifested themselves while using python?

to state my question more precisely: which language best forces you to get healthier programming habits?

[winter spear]

Static typing isn't inherintly better, it's just harder to shoot yourself in the for

[buoyant zealot]

I mean that's kind of what I mean, I'm really lazy regarding any defensive programming habits or stuff like that, I thought it could be a decent idea to learn sth else that is a little more strict.

[fallow terrace]

not having type declarations is not a bad habit as long as you take the proper precautions. you can use a static type checker if you want, most will have a decent type inference engine. what are the others you're thinking of?

not requiring type annotations is actually rather good for teaching, imo, because types can be a fairly complicated concept

[buoyant zealot]

yeah no for learning how to program it is nice, but if longer code is written or someone else needs to read the code, type declarations can help

[fallow terrace]

which language best forces you to get healthier programming habits?
this question cannot be answered until you actually define what you mean by "healthy." decide which habits you think are healthy, and choose a language to teach that has those.

[fallow terrace]

ehhhhhh. most editors can have some way of presenting the inferred types to you, so it's fairly easy to read.

and fyi, this channel is about teaching programming, so just programming in general is out of scope

[buoyant zealot]

yeah, but like, if I'm not forced to I usually just forget about it until it messes me up. another bad habit I have is sometimes space allocation or just not using the most optimal way to deal with arrays because it is sometimes shorter to write or just works, which can sometimes reduce efficiency since array concatenation etc. or copying without getting some pass by reference error can be time consuming

[buoyant zealot]

oh sorry, yeah what would be the better channel for sth like that then?

[fallow terrace]

perhaps off topic, if you're also discussing non-python languages

[buoyant zealot]

but these channels have all... weird names?

[fallow terrace]

!off-topic-names (these ones)

[worldly dew]

Off-topic channels

There are three off-topic channels:
• #ot0-psvm’s-eternal-disapproval
• #ot1-perplexing-regexing
• #ot2-never-nester’s-nightmare

The channel names change every night at midnight UTC and are often fun meta references to jokes or conversations that happened on the server.

See our off-topic etiquette page for more guidance on how the channels should be used.

[buoyant zealot]

oh🐔

[noble aurora]

i disagree. it's very easy for someone learning python to not know what type some variable is, which is a Bad Thing. with a language like java, this problem doesn't exist, since you need to think about what type some variable should be

a common one is something like

N = input()
if N > 3:
  ...

this still errors, but it's hard to figure out why the error happened, even if you understand what the error is saying

[fallow terrace]

hmm, that's true.

[viscid wing]

Heya, so when I write in a file with my program it replaces the last text, how can I make it so that it just goes to a new line?
This is the code: py else: print("Aircraft added successfully! Now returning!") f = open('airspace.txt', 'w') f.write(f'Callsign> {callsign}\nFlight level> {fl}\nLanding runway> {landing_runway}\nNotes> {notes}')

[viscid wing]

anybody

got any idea?

[desert patio]

yeah i got an idea that's off-topic for this channel

[viscid wing]

oh crap didn't realise I was in this channel 🤣
Thanks for letting me know

[pure fog]

to be fair, a lot of newer statically-typed languages have good type inference, and sometimes explicitly writing too many type annotations is even considered a code smell

so i think it depends in part on what bad habit you have in mind

i think any statically typed language will help a student break the habit of writing "ridiculously polymorphic" functions

otherwise i can't think of many bad habits that are specific to python and can be unlearned with other languages

[misty dirge]

Do you have an example of a ridiculously polymorphic function?

My observation is that people who haven't wrapped their mind around types tend to have a hard time breaking things down into functions in the first place.

[dull cave]

For what reason should it be a problem ?

Oh I see your example, ok.

[dull cave]

Do you think that people should start with java for example for this reason ?

[noble aurora]

no

[bleak parcel]

Hey guys, does anyone has ressources to start developing network scanner ? Not ressources on "Code ur python network scanner in 30mn" more something like "how do you use network protocol and the os along python to create a network scanner" ?

[dull cave]

Oh ok

[wooden grove]

I think there's actually a similar problem in teaching math, because the types there are also implicit.

What I mean is, students get confused when they don't understand that there are different types of mathematical object (number, function, set, random variable, etc.).

[noble aurora]

interesting. the problem can't be that bad though, right? even if they're not using the correct notation, they still know what they want to do with it. you're not going to get a type error on paper. maybe i'm misunderstanding though?

[wooden grove]

Yeah I mean its more just that they fail to really understand what's going on.

[dull cave]

That's an interesting point of view

[wooden grove]

E.g. the expression x y can mean different things depending on the types of x and y. If they're numbers, it's multiplication. If they're matrices, it's matrix multiplication. If x is a function, then it can denote function application. So we stick to conventions that make it clearer, like always giving functions names like f and g, using uppercase letters for matrices, and always wrapping the arguments in parentheses in the case of function application x(y).

I can't think of any good examples of the kind of errors students make 🤔

But I know it did cause confusion for me personally.

[gloomy carbon]

use a+ instead of w

[viscid wing]

Yeah, thanks it worked

[gritty lintel]

hi..iam new here idk what we do in this server

[ripe folio]

I know my classmates keep mistaking expressions for equations

[hidden oyster]

Hello! I've been writing Python code for a couple years now, and would actually like to find someone to review my coding practices. Does anyone know where I could find someone to do that? (paid) (Please tag me when replying)

[misty dirge]

we can't advise you on finding someone you can pay to review your code, but you can request code reviews in our help channels. see #❓｜how-to-get-help

[edgy trail]

@hidden oyster yes

[misty dirge]

yes what?

[edgy trail]

He should request reviews in your help channels

[hidden oyster]

I'm not looking to share any code publicly to get reviewed. This is proprietary code from my business. So ideally I find someone who can sign an NDA and invoice my business

[misty dirge]

we can't help you find that on this server. sorry.

[hidden oyster]

Thats a bummer. Thank you though!

[misty dirge]

@icy marlin this is not a help channel; see #❓｜how-to-get-help

[misty dirge]

no rush, but I'm still interested in your thoughts on this

[vale current]

I guess you can have somewhat polymorphic functions in statically typed languages too? But they aren't "ridiculous"

[misty dirge]

even in Python, the most "ridiculously polymorphic" function is still going to fail if it doesn't depend on some meaningful shared property (ie duck typing)

[vale current]

I don't think I've ever had a problem with writing functions which might be considered "too" polymorphic, they've typically always served the purpose of working on some a somewhat specific type of input

[native hearth]

Examples aside, what does "ridiculously polymorphic" mean

[vale current]

Maybe a function which tries to adapt to every type under the sun?

[native hearth]

so print then, first thing that comes to mind

[fleet pulsar]

print can actually just be typed with Any/object

I'm thinking more about functions that return different types depending on some property of input

Like this, for example

def funcy_stuff(items: list[int]):
    if len(items) == 1:
        return items[0] * 5
    return [i * 5 for i in items]

[fallow terrace]

can't that be annotated with an anonymous sum type?

sorry, union

[fleet pulsar]

That would be typeable, but neither statically verifiable nor obvious to the average programmer

If I pass lists of items and receive lists back, I'd be surprised when my program crashes just because the list had a length of 1 and it decided to return the item itself, without a container wrapping it

[fallow terrace]

ah I see

yeah unions are kind of a nightmare. I suppose that's why most statically typed languages (that I know of) don't have them

I suppose you could do some weird stuff with type(), but that sounds like pain in a statically analyzed program

[fleet pulsar]

Unions as a concept are pretty good, especially tagged unions (rust enums <3), but can lead to all sorts of issues when [you assume the type, and the union itself has very loose semantics]

IMO, the problem here is specifically the function using very specific and unclear conditions to juggle around return types, input types, behavior, etc

Because really it's just two different functions packed into one because (from personal experience) someone didn't want to unpack the result when they were sure there's only one item, so they specialcased it

Two overloads (funcy_stuff(list[int]) -> list[int] && funcy_stuff(int) -> int) would be fine because you know what you're getting yourself into, but the way this example function is currently written, it's just asking for trouble

That would be the most obvious example of ridiculously polymorphic functions that python... I wouldn't say encourages, but fails to prevent

[fallow terrace]

To be clear, I'm specifically referring to unions as anonymous sum types. Named ones are a different matter altogether.

[fleet pulsar]

Ah, fair enough

[orchid thistle]

imo sum types are reasonable but "unions" as in "if an A ∪ B happens to be an A you can use it as an A" isn't really statically typed at all

[fleet pulsar]

Yeah, exactly

[orchid thistle]

either you say that the value does in fact have a runtime type that isn't in its static type, or you allow reinterpretations between any two types which is just weird (unsafeCoerce)

[fleet pulsar]

Although you can type this function as -> int | list[int] and mypy wouldn't complain, you're still changing the return type on a whim, deciding on an input's property (length of the list) rather than type

Essentially just asking for trouble and tough to debug issues

[orchid thistle]

changing the return type based on the argument is fine as long as how you do that is in the signature

(i.e. if the language has proper dependent typing)

[fleet pulsar]

True

[orchid thistle]

if not then yeah just don't do that

[native hearth]

what do you guys do when you're trying to help someone stubborn with python
what's your go-to technique? do you reset and start from the top?

[green warren]

Stubborn how?

[native hearth]

insisting on doing things a certain way, basically overcomplicating simple things they already know how to do

[green warren]

Can you give an example?

[native hearth]

im asking in general really but #help-pancakes is the most recent concrete example
it worked when i reset things and took it from the top to break down the problem and explain how to solve it

but I think I felt even more frustrated than the person having the problem until I reset

[green warren]

I'm not following what's "stubborn" about them, I think they just didn't know how things work.

Also, doesn't str.split() (with default arguments) split on all whitespace, treating multiple spaces as a single delimiter? That seems like the simplest approach.

[native hearth]

in this particular case they insisted on the importance of whitespace even when it was demonstrated that it didnt matter, they wanted to do things a certain way based on some misunderstanding of the test cases

maybe not the best example of stubbornness