#general-chat
1 messages ยท Page 136 of 1
@sick adder shirt for Hacktoberfest in October? This year?
@vestal phoenix yup!
I wish there was an accurate easy to use google maps of internet infrastructure
123net provides one for their fiber, and I love it, but it's almost never this easy to find and most of the time we care more about intersections than exactl locations since you can't just tap off a fiber mid way, you need all sorts of equipment and the fiber probably wasn't designed for that use anyway
woah RISK5 is pretty cool! https://www.youtube.com/watch?v=jNnCok1H3-g
Thanks to SiFive for sponsoring this video! Check our their core designer at https://www.sifive.com/core-designer We get a look at some working Open Source s...
Battery problems forever solved. https://www.androidpolice.com/2018/12/05/unlimited-power-asus-rog-phone-charges-when-plugged-into-itself-solves-worlds-battery-problems/
lool
One time I got so stressed about something I got very sick
I never knew stress could have those effects
@jaunty jetty I work at a hosting company in Lansing so I just stay after my shift, get what I need done, and use mobile hotspot for the rest
I've given up on the idea of getting fiber though, there's nothing affordable out there
I just want sweet graphs
would 'you' buy a featherwing board that's an I2C expander and then a bunch of 3 pin headers with jumpers, so you can configure settings in hardware and then read them on an I2C address
what if instead of 3 pin headers it was a whole bunch of dipswitches
like a lot of dip switches
Those little rotary switches are also cool for setting that kinda stuff.
@dusty citrus yea, I've been following a CNC etch a sketch that looks tempting to build.
You say that as if there are many
@karmic kite Just saw this, after a sweet nap . So far I plan on building my CNC router without any plans for the base or table, just gonna wing it and learn as I go.
I saw a CNC etch a sketch in one of the youtube videos I've seen. Since I'm new to electronics, my two new favorite subscriptions are to "bigclive" and "greatscott"
am I the only one that sits with things in their adafruit cart wondering if there's anything else I should order?
I usually go till a large order with freebies:P
Yea that etch a sketch on hackaday looked pretty neat
@cunning shuttle That's what I use Wishlists for
I keep a couple wishlists -- a few for specific projects, general interest items, and consumables (like headers)
not a bad idea
When I get a more urgent need for a project -- I wait until a broadcast (coupon code), and get enough for at least free shipping (or to justify the shipping cost)
yeah, i had stuff in my cart last night with coupon code, forgot to hit submit so added it to a wishlist until JP's broadcast
You have his code for today? If not -- "lowpass" is good until midnight-ish.
yeah, thanks, I watched the broadcast ๐
only $96.40 more to get a circuit playround ๐ค
maybe next time
not saying its broken but if you put the coupon on your cart... you can wait a few days before pressing submit
Sometimes the powers-that-be....cough....forget to take down coupon codes sometimes. Hey, they get busy with life, etc. At midnight I'm sometimes out like a light, so I don't fault them.
It works out for everyone -- you get a discount and they get a sale.
I'm not complaining by any means ๐ I spend a pretty penny either way
my only complaint is the tarriffs for international exchange
any chance one've yalls has at some point been inside a launchpad style device?
looking at taredowns they look a lot like just really extra big neotrellises
can't find an image good enough to read the ICs
the teardown's closeup isn't very close http://www.illuminatedsounds.com/wp-content/uploads/2009/11/71.jpg
yea that's just a crop of a photo from another site
the full size is tantalizingly close to being readable
The sickest VFD russian clock out there! the big ELEKTRONIKA 7. an absolute beauty and from what i've heard were all over russia. If you have seen these befo...
I'm seeing packs of these IV-26 VFD tubes, and IV25s which seem pretty similar, handfuls for under 20$
he finds out that this clock is running the tubes at 11v, 30 would be better, but both of those should be pretty easy to drive from an M0
well, control from an M0
Are you computers really divided into PCs, Minicomputers, Mainframe, Workstations and Supercomputers ?
Like all those terms are they still used ?
@jaunty jetty
I feel like there's a conversation prior to this that I'm not aware of
minicomputer is not a term but there is 'small form factor', some proprietary formfactors like the nuc, and ITX
mainframe is a specific kind of machine architecture and yes they're still used, they are suitable to certain kinds of tasks like payment processing
workstations aren't a specific thing other than a station used primarily for work; in the abstract workstation implies that the work to be done is of a load that generic computers couldn't effectively support. But there are laptop work stations and desktop work stations and the imac pro and so on
Like a Computer just meant for a single purpose ? (In context to Mainframe)
it's more topological than that
you use a mainframe when A, you must use a single 'computer', and B, you need a huge amount of processing and hardware
mainframes tend to use processors on expansion boards so they can have a very large number of them, and they're attached to data rooms full of SAN, with double digit high bandwidth NICs
and of course, they run specialty OS
Mmmm They are like High Powered Computer used by only a single user ?
with payment processing every time you involve an additional 'computer', it gets significantly harder to maintain data integrity across the system, so you need the smallest number of machines handling the work, and each machine needs to operate at an absurd level
Mmmm I see.
Workstation in it we reffered to as drafting grade desktops
Drafting Grade ?
I'm going to guess that Visa has fewer than a hundred mainframes handling all of of the transactions in america
if you had a company, and some of the employees of the company needed to run CAD software, those employees could not work effectively with hardware comparable to what HR uses
Some businesses have a mainframe for visa interact and Mc on site. Highly secured of coarse
So When you can't have Servers as it gets complex to maintain data across them you use Mainframe ?
right
Perfect Makes sense.
and when you need employees to have a lot of power to do their work, you give them a workstation class device
Instead of Low Performance Large Number of processors you have Small Number of High Powered Computers AKA Mainframes.
Aaaah workstations make sense.
PCs with Redundant Memory and High Performance Hardware ? (In context to Workstations)
at my last job I 'exclusively' used a macbook air, 2011, which is an impressive computer but the job consisted mostly of looking at websites
not redundent, but probably ECC
redundent memory is not so much a thing
memory is extremely expensive and workstations tend to have a lot of it
you don't raid memory, you raid storage
the workstation shouldn't be responsible for storing data, that should be on a server, for performance, stability, and security
if anything a workstation is likely to have RAID to get better IO performance not redundency
Mmm Faster Storage ?
but even then, a decent SSD will beat a raid... 0? I think it's 0, and SSDs in raid are better but not in a useful way
I'm not sure any motherboards currently support multiple NVMe devices
Mmm They do
Generally the High End AMD Ones
Not sure if they support RAID
BTW What do you think about the Computer Generations ?
I don't
there's always software raid, you can do software raid on any set of mounted drives
the thing is SSDs are already so fast that most applications are not capable of getting useful benefit out of NVMe to start with
CAD is the typical example of a heavy workload and it just doesn't involve that much disk access
video editing does, and some very high end audio editing, but video editing is like three exponents larger data so it's a bigger deal
Mmm Makes sense
๐ค
Hey One More Doubt
Whats like the difference between Cache and RAM ?
ok so, you've got your house, you've got the store, you've got the store's distribution center, the processing center, the farm
you can just go eat an apple in your house
you have to go to the store to get an apple if you don't have one in your house
Yup
if the store doesn't have one, it's gotta come from the DC
OFC
etc etc, tree gotta grow apples and that's slow
L1 is very very fast and very very close to the actual transistors that do stuff
L2 is not quite so fast but still faster than ram, and a bit further away
L3 is about as fast as ram but relatively quite close and because it's sized in megabytes the access speed is faster than gigabytes of ram
ram is much faster than SSDs, and it's like two inches away
Oooo I see
in that same order, L1 is ludicrisly expensive
RAM is so expensive right now that it's influencing the industry
L1 is more expensive
So How does the Processor decide where the Memory gets stored whether in L1 L2 L3 or Ram ?
it all funnels down
L1 is what the processor is using now, L2 is what it thinks it will need soon, L3 etc etc
Aaaaah I see
caching logic has a huge impact on performance
?
an, sure
so then supercomputers are kind of a non-term
supercomputers are real fast, usually designed for doing computation in a specific domain, they tend to contain mainframes and are almost always not a computer
most supercomputers are networks of discrete devices but somebody for marketing reasons decided to call it a supercomputer
Hey So One doubt aren't Supercomputers like Many Mainframes ?
And Not always computer ????
pretty much all of them use cluster topology, but then they're also going to be hybrid topologies because clusters aren't good at all workloads, sometimes a mainframe with five thousand GPUs is going to be faster and more efficient and less expensive but you also need cluster nodes and once you have those you're going to need the SAN to be it's own device with it's own topology and all the different topologies talk to eachother over very very very fast networking
SAN ?
if google decided to label the cloud that runs youtube as a supercomputer, they'd be right
Storage Area Network
BTW Mainframes can have 5000 GPUS ???
probably
I think Canada just built one that has dozens and dozens of racks, and each rack is full of 2U chassis, and each 2U chassis has like 5 nVidia Titans
Mmmm So High Performance Computers meant for a Specific purpose all tied together are known as super computers ?
Titans ?
๐ฑ
computers that have been designated as supercomputers by their owner are known as supercomputers
These Computers are Beast.
it's kind of like mid tower, tower, full tower, extended; the words don't mean anything and a lot of companies do whatever
Mmmm I see.
btw if you want to build a supercomputer, it's fairly approachable to put together a beowulf cluster
this is what cool nerds did before processors with multiple cores
looks like the board contains individual power supplies and ethernet adapters for each, plus it exposes an OTG USB A port
doesn't look from this article like the board is available for purchase, which would be a sad truth
it must be quite expensive, but I'd buy a 4 position one if I could
also if they're doing real computation on this they're going to need heatsinks, they're going to need at least one fan, and they should add 3D printed shrounding
๐ฑ
screams, all his family members give him an wierd look
awkward
I thought Parallel computing could be only done by high end parts.
Thats sooo cool Raspberry 0s in Parallel ???
you can usually parallelize in software
the thing with mainframes is they parallelize in hardware which has a lot of security and scale implications
Mmmm.
the overhead of software would cause there to be a maximum viable size of a pi cluster based on the network hardware and the master controller, like even for 16 pi 0 you probably want the master to be at least a pi 3B+
So If I had two computers could I like parallelise them too ?
Ooooh so we need something like a Master too ?
This is really freaking cool.
But sadly It seems like I can't just use the same Windows on em
yea, so the typical implimentation of a cluster is you've got a higher end machine with very fast networking that's connected on one side to a switch, and the other side to the rest of your system, behind the switch is the cluster. Your system sends 'tasks' to the master, it distributes tasks among the cluster, and either the cluster returns completed work to the master or the master acts as a router so the nodes can send completed work back directly
the systems beyond the master don't have direct access to the nodes at all
this makes the master a bottleneck, if it's not fast enough to handle ingressing and egressing tasks, then the cluster can't reach full utilization and will be inefficient
cloud infrastructure can be pretty similar but you tend to have clusters of dedicated task servers, and the clusters can interact with eachother in layers
Oh I see.
http request comes in, hits the router, router sends it to an API node, API node figures out what's being requested, sends tasks to webhost node, compute node, san node, they return their tasks and it replies to the http request with a compiled document
Mmm
it can get a lot more complicated because with routing trickery the reply doesn't need to originate from the API node, so it could be that it sends tasks to the compute node and san node which then send their work to the webhost node that compiles it with fixed html stuff and replies to the http request
Okay So API is Master and other nodes are like rest of the cluster ?
So How do the various data from different nodes get compiled into one request ?
Threads are a wonderful and terrible dark hole
that is the entire field of cloud infrastructure
for a lot of things it doesn't, you don't compile images into html, but you embed in the html the url to load the image, so the requestor then sends another request for the image, and a lot of cloud computing relies on async and callback functions, and separating requests that return a payload from requests that invoke an action
Oh I see....
Kinda makes sense.
Mmm The ALU is kinda confusing (You know the ALU in a CPU)
there's a reason cloud engineer is one of the legitimate engineers
Mmmm Yeah True.
ALUs are out of my knowledgebase
I just know the fact a CPUs work this way:
CPU divided into Memory CU and ALU
CU Reads Memory extracts and decodes the data and acts accordingly. It passes on the data to ALU and ALU Accordingly solves the problems like a Good Calculator. And Then CU takes the data and stores it back in Memory
yanno, since the ESP32 has a core dedicated to the wifi, it'd be suitable for making a cluster, until you break 2.4GHz in half like a twig
Break ??
wifi congestion compounds fast
if you had 200 ESP32s in a small area, it would be very difficult to get anything over wifi
๐ค
you could do that
but SAMD51s are single core and I'm not sure how the USB interface works so the device might be out of contact any time they're doing work
Oh I see
You need like Dual cores ? Where ones active on network and others for work ?
that said I like the idea of a board with a USB hub and a bunch of microcontrollers
Wait why would you do that ?
you don't need it if you have an independent network device that can interrupt, I really don't know a lot of things here
Well TBH You do know a lot of things
if you were going to make a compute cluster using M4s, buying a bunch of any existing M4 board would be crazy expensive, and potentially have a lot of features that aren't useful like large amounts of pins and battery controllers
if you put those M4s on a board, you'll probably pay about 2$ for each
BTW Didn't know this LOL
Agreeed regarding the M4s
Dude. One day I'll make this come true. 10 M4s on a single board using Parallel Computing.
But Long time for that ๐
you can probably get a 4 client USB hub IC, maybe 8 but I think it's usually 4 at a time so you'd chain them and get 3 clients out of all but the last IC
if you put 4 hubs on a hub, then you'd have 16 end clients, and 16 M4s would be pretty fun
But Writing the code for em would be harder
21 ICs for that, then a bunch of resistors capacitors and power regulation
How would you divide and give em instructions ?
oh no, so far these are just a bunch of M4s
literally as if you'd plugged 16 M4 feathers into a bunch of USB ports at the same time
oh and we need a flash chip for each, so that's 16 more ICs
Why flash chip ??
circuitpython
And How 21 ICs ?
4 Hub ICs with 3 M4s each. So thats a 12 M4s and 4 USB Hubs right ?
And Mmm I see CP ? You mean the drag and drop ones ?
5 hubs, one hosting 4, each of those hosting 4 M4s
then the M4s should have a way to talk to eachother without going upstream, just in case that's useful, and I'd want to expose as much of the remaining SERCOMs as is cheap
Oh I see
So you same USB Bus for Connection between the M4s and also the Host ?
๐ค
no, definitely not USB
So Use SPI or I2C ?
maybe but they're master/slave protocols
Okay Jeez
๐ ๐
No other Protocol ?
What about Serial ?
Does that even work between multiple devices ?
๐ค
serial would work but RS232, if that's the one, doesn't have addressing
I think there's a later serial that does, maybe there's support
design your own protocol?
why reinvent the wheel, surely there exists a well designed protocol
just keep in mind that apps say words that don't necessarily relate to reality
nobody talks about programs in terms of generations, at all
they're talked about in terms of how they handle types, how abstracted they are, if they're functional, imperative, etc and to what extent, if they run natively or in a VM, etc, etc, etc, but generations just isn't a thing
more accurate would be abstractions like the OSI model
Then why the hell does it need to be taught ???
C is a very low level language, it has very few abstractions and runs natively. it is designed from the ground up to be imperative and doesn't have a lot of support for types outside of the ones it comes with
the OSI model is for data systems like networks
๐ค
it comes up in a few areas, but yes around things that talk to each other
most programming languages can't be usefully described on the OSI model
๐ internally thinking What are they talking about?
you could have an OSI like model with different levels of consideration that would work relly well for programming
so, at the physical layer, SPI uses 3* wires, one is a clock, and the other two change voltage levels in time with the clock to transmit data
Okay Just a way of Characteraizing things ?
at a data layer SPI transmits bytes...? I think
By things Communication Links ?
at the network layer SPI uses a Chip Select pin and a master/slaves topology with slaves rarely if ever talking to eachother
the transport layer doesn't really apply because it's a single collision domain many to one topology, you don't have subnets or routing between nets
that's talking to the master
at the data layer what matters is who is transmitting and who is receiving, having the master communicate messages from one slave to another would be transport layer behavior
session layer also doesn't really apply to SPI because it's just, is there power
Mmmm Oh-kay
presentation layer is the SERCOM maybe, or the code that interprets bits into meaning anything
so, this is all logical abstraction to break down the problem of making an SPI that actually works and is useful for things
you have to be able to transmit electrical signals as much as you have to be able to send and recieve data in your program
in programming languages, no matter what at some point instructions have to be executed on the processor, but there's a lot of stuff between a window and there
so like I said before, C is a very low level language, the instructions that execute on the processor are highly reflective of the code you wrote, while in C#, the code you wrote is part of a very large system and there are multiple layers at which 'the machine' takes your code as a suggestion for what needs to be completed rather than direct commands about what to do
Lemme be honest I don't get half of it. ๐
Mmm Last two paragraphs make sense tho
you don't have to understand the OSI model (I don't actually), but you can tell with a little understanding that it's actually a useful model, where numbered generations with vague implications about complexity and abstraction are not
if you're going into the networking field you will have to know the OSI model
I just get the fact its some kinda wierd way of Describing protocols and their characteristics
I dont know where am I going in future
RN I am trying to learn the basics and get into Web dev to get into Proper Programming and earn money ๐
But I am still super interested into ML and AI
that's fine
it's complicated
it happens when you've been an adult too long
Oh-kay.
an industrial electronics supplier near me is hiring for on-camera talent
and the listing for on camera tallent has one-click apply
but the listing for data entry requires you to send them an email
Talent ?
in the showbiz sense
learn the basics and get into Web dev to get into Proper Programming what?
web developers do proper programming (depending on your definition of proper). It's just often a different language
also depending on your definition of web developers
The thing to remember about the OSI model (aka standard X.200) is that it was originally developed by the ISO and CCITT/ITU-T and the only set of protocols that are really expected to confirm to it are the various X.? protocols.
It can be useful as an abstract idea of how various network protocols work together, but it is not definitive and the TCP/IP stack does not really map 100% to it
the big takeaway from the OSI model is that different layers in the stack do different things, so that you can swap out one small piece without the rest of the stack knowing or caring. Hence why we can run TCP/IP over so many different physical network layers.
Staying in downtown Manhattan on the 20th for 3 days
one day of ad blockers
god i hate stuart little, he stole my wife and kids
@cunning shuttle
I meant I'll get into Web Dev. So That I become a Proper Programmer
I didnt mean Web Dev aint Proper Programmers.
Thanks ๐
step one, learn HTML, XHTML, HTML5, CSS, XML, JSON, SQL, JavaScript, and photoshop
step two, learn Node.js, linux, and get familiar with VMs
step zero: don't learn PHP, not even a little
what do you have against PHP?
it's an incredibly poorly designed language
also it's over, node.js is the future, apache isn't, php isn't, even mysql is not all that much. becoming a PHP developer puts you into this dying class of web developers that will continue to exist for the exclusive purpose of maintaining wordpress and joomla installations
I used to know a mechanic who stopped being a mechanic because he didn't feel like learning how to fix modern cars with all their electronics and computers; learning PHP now is like becoming a mechanic that only fixes cars that don't have engine control computers
oh and it will make you worse at all other programming
it's that bad
I can't say I agree, but to each their own
I have a similar opinion of JS: it makes people worse programmers.
yea, and if it weren't for node.js and html5 being deeply entangled with javascript, I would say steer clear
that's (node.js) && (html5 being entangled)
My sites are generally HTML 3.2, with application server backends.
one of the bigger actual problems with php is that it's not suitable for cloud hosted active content websites, which is partly intrinsic and partly that php is mostly dependent on webhost software that is deeply incompatible with cloud hosted websites, like apache, and this is a lot of why NGineX...NG... what's this name... NGINX, it's pronounced engine x, but it's spelled n ginks, because this is the future and everything has to be unclear
nginks is basically an all purpose API node, it gets http requests and either returns flat files or forwards them to other software, like maybe node.js or a database
side note, huge advantage, if you use node.js in your web pages, and you use node.js in your backend, then hypothetically, you can use node.js everywhere, and that's really valuable. You don't have to be writing the client half of what is functionally a single system in one language while you write the server half in another, and then maybe the mobile app in a third and fourth. Even if node.js is build on javascript and therefore has a lot of intrinsic problems, being able to use the same solutions for the same problems has huge knock on effects in improving code quality and reducing maintenance effort
this was a big selling point for Swift, Apple's fancy new language that was supposed to replace Objective C and would be crossplatform (a JVM/CLR type system) so you could write the servers that support your iphone apps in swift and not just do the above but also share a lot of actual code between them. I think that's stil a real good idea but the people I follow who were really into Swift have gotten significantly less optimistic over the last 6 months
I just applied for 3 cobol positions at 3 companies, some dead languages never die.
yea.... banks.... banks.
that oil pipeline from alaska to the rest of america... still runs on code written in the seventies and eighties
SCADA is still really popular around here as well
SCADA? I've heard of SCALA
https://en.wikipedia.org/wiki/SCADA more architecture then language
Supervisory control and data acquisition (SCADA) is a control system architecture that uses computers, networked data communications and graphical user interfaces for high-level process supervisory management, but uses other peripheral devices such as programmable logic contr...
yea this looks like smarthome stuff, for industry
ancient
XML is just JSON but different, XHTML you can probably skip yea
Mmm XHTML is for Apps or something ?
Photoshop, or GIMP but photoshop is much more better than it used to be and GIMP is much less good than it used to be, is super useful in general, and especially so if you want to make a website that looks good but don't have a graphics designer to send all the graphics design work to
XHTML is for websites
it came between standard HTML versions and HTML5, and it's basically HTML but XML compliant where normal HTML can't be defined well as a syntax due to it's diverse special cases, indefinite nesting, and general lack of consistency. HTML5 mostly makes XHTML redundent
Mmm I see
on the topic of out dated languages. why am I diving into erlang
just reading the section above the contents table on wikipedia, because it sounds interesting?
going to build a couple of IP servers here on erlang as poc
it seems slightly more effecient then java
in this context, what does that graph mean?
kinda like alot of interrupts on a mcu are bad, non voluntary context switches stop other threads to take precedence
its saying over millions of threads and sub processes theres constant "interrupts"
so XHTML is defunct because what makes it "better" than the contemporary mainstream HTML version wasn't that big a deal to most people, and HTML5 is distinctly superior in a lot of ways, like how it supports streaming video players without plugins
... so what your saying is HTML5 is flash-ier then HTML?
looking at wikipedia's example code for Erlang, which is still being updated (15 days ago), it's not a big deal language because it's too disimilar from english
actually, the opposite
a lot of the drive to develop the HTML5 spec was that everybody hated Flash and wanted to make it defunct
give me some credit it was for the sake of the pun:P
game servers seem to love [erlang]
I bet, these features sound amazing for that kind of application
haskel would be good for it too; yanno facebook uses haskel to code it's spam filters
but if you get too far from things that people can just look at and vaguely grasp, you see a rapid decline in success
I love coding in haskell, its a fun logic system to play with. I just never found a practical use for it
-module(fact). % This is the file 'fact.erl', the module and the filename must match
-export([fac/1]). % This exports the function 'fac' of arity 1 (1 parameter, no type, no name)
fac(0) -> 1; % If 0, then return 1, otherwise (note the semicolon ; meaning 'else')
fac(N) when N > 0, is_integer(N) -> N * fac(N-1).
% Recursively determine, then return the result
% (note the period . meaning 'endif' or 'function end')
%% This function will crash if anything other than a nonnegative integer is given.
%% It illustrates the "Let it crash" philosophy of Erlang.```
heeeeey
-module(httpclient).
-compile(export_all).
-import(lists, [reverse/1]).
nano_get_url() ->
nano_get_url("www.google.com").
nano_get_url(Host) ->
{ok,Socket} = gen_tcp:connect(Host,80,[binary, {packet,0}]),
ok = gen_tcp:send(Socket, "GET / HTTP/1.0\r\n\r\n"),
receive_data(Socket, []).
receive_data(Socket, SoFar) ->
receive
{tcp,Socket,Bin} ->
receive_data(Socket, [Bin|SoFar]);
{tcp_closed,Socket} ->
list_to_binary(reverse(SoFar))
end.
in my code example, nothing the comments describe would have been intuitive to me
low base intuition trends with low adoption
export is public and private, the functions I'm trying to get used to
module and import arnt that far out there
I could probably guess -module(etc), but the . stumps me
;
. is end of statement
and then we're getting arrays of divided references, .
what intuation (of mine) appears to apply is immediately wrong and then later provided conflicting evidence
put the result of fac(1) into object 1
the line after that is by every other language I've ever looked at, utter gibberish
are you on wiki or?
just that block of code I quoted
ohhh I get it
foo(int)->{io:format("This is an integer\n")};
foo(string)->{io:format("this is a string\n")}.
there one function with multiple inputs and overloading
so my understanding of the example I started with, is that there's no actual if statement, the fac(0) line is an overload that only accepts calls where arg1 has a value of 0, and in all other cases it goes to the next defined overload, which accepts all other values and crashes on negative values
is it actually an if?
if it is just highly specific overloads, which is cool, then what actually does ; do
can't have an else without an if
again weird language
so is ; for demoting the not last in a series of overloads?
mmkay
this is making me want to go back to scala
does using a pi as a proxy to do things to your traffic have performance consequences as a result of the pi not being that fast or is the load small enough?
full on proxy or dns server?
I want to block twitter after accessing twitter.com/ a certain number of times, since the computer will cache DNS requests, I think it will need to be a proxy
accessing twitter.com/ will only happen when I open a new tab or return to the timeline, other accesses won't be counted
I found the pi bottle necked my bandwidth, now I just have a dedicated box with gigabit I use for reverse proxy and web hosting
although for blocking twitter and fb I use opendns.com
I don't want to block twitter, I want to cap it
if I go to twitter 20 times, the 21st time should fail, so that I'm not spending all day mindlessly opening it
Ah I see
but if I'm just browsing twitter it shouldn't count
you'll need to log sessions not just dns inquiries
https://www.unix.com/man-page/All/1m/gated might be useful
requests for the root directory only occur when the page loads on the timeline, so I can just count those, but if the pi isn't suitable for being a proxy, I don't and won't have a better option any time soon
are you a heavy bandwidth user?
updating the timeline is ajax requests to webhooks, and other pages or specific tweets aren't teh doot directory
I can be
opening a few dozen webcomics at once tends to lag firefox
fair
at worst, you can try it. If it limits you as much as it limited me you can be prepped for having a dedicated box
Looking at my totes of components. Feel like a hoarder and kinda want to throw them all away
I;m thinking next repair cafe in the city i might donate them all
Merge a pen with a calculator and you get the CalcuPen, a barely useable yet highly collectable piece of 1970s tech. NOS LCD Pen Watches on eBay https://ebay...
it's not the best pen design, mostly the nose area ends up as a large conic angle that can interfere with viable writing grips. I'd love to take one of these and add a bluetooth module to use it as a phone control stick
he also talks a bit about watch pens, and I'd love to makea watch pen that uses radial lightguide bands and binary time
using a thin long PCB with side throw LEDs, a row on each long side with the rows offset by half so the rings can be closer together, probably charlieplexed 4 bits for the hour, 5 for two minute group because who needs minute precision from a pen. if only I had a lathe to cut the pen body on
no makerspace/work shop near you?
detroit
nearest one is 15ish miles away, 25 miles by road, and 50$ per month minimum 1 year up front
Doesn't seem that far away
it's like five cities away
interesting.. different world views on distance I guess
this is how the area is organized http://gppreview.com/wp-content/uploads/2014/08/The-Real-Sprawl-Problem-e1408331139594.png
I think this place is either in the city of detroit or dearborn, and I'm a ways away
round trip would cost about 16.80$
detroit problems, measuring distance in dollars
fair enough, I've toyed with the idea of making mini lathes
make a levelling mount for a die grinder, couple straight edges
man the guy who wrote this library has some issues with adafruit...
looks more like adafruit has various outdated things, and the guy has provided detailed unemotional help on how to work around that
given he's not affiliated with adafruit this seems well beyond his responsibilities
This was one of the first libraries I used for a display... kinda want to go back to this project again
Was making a call of duty capture the flag style paint ball thing. If you could stand in the area for so long you would "capture" it, and be able to move to the next post. Didn't make it far back then but I've learned enough I might be able to get it going again
Thinking I'll swap the wifi for LoRa though, just to get across the full field
Wooooo hooo!!! My adabox is gonna arrive on wednesday!!!
Hey @viscid folio might I make a recommendation for a modification to the Trellis M4 learning guide?
wazup
For some reason I did not expect that as a response from you hahaha.
anywho, In the learning guide you mention that we can use FastLED to control the neopixels on it. At the moment FastLED does not support the M4 boards like the Feather M4 and ItsyBitsy M4. I'd like to recommend adding in that FastLED is not yet supported on the M4 chips. Unless of course it is actually working on the Trellis M4?
oooh shoot
good point
i dont know that it is
do ya wanna do a PR to fast LED? ๐
I would definitely do a PR to FastLED and make them work however I dont currently know how to add in the support for the Trellis, Feather and ItsyBitsy M4 boards. Someone attempted to add in basic support for the Itsy Bitsy M4 and FastLED merged it into the main but it does not seem to compile for the ItsyBitsy M4
Alrighty
oki done
Im hoping that someone gets FastLED working for the M4s soon. My lightsaber project depends on it.
And cool beans!
somebody aught'a put in a guide for how to add support, it should be just a process of establishing configurations, but I've been watching this whole thing you did that
just as a note, FastLED doesn't have a folder for SAMD51 at all, the one M4 configuration block was added in SAMD21, so the whole thing needs to be refactored
Yeah I saw that @jaunty jetty
that might be part of why it's not working
I believe there is a guide for porting FastLED to other boards. But its much lower level than I currently understand.
This is the guide I was thinking of @jaunty jetty http://fastled.io/docs/3.1/md__p_o_r_t_i_n_g.html
this could be a .md file in the source
also this is assuming a whole lot of familiarity with the project, to the point of being meaningless
Yeahhhhhh
it does say you should make a new folder for the family, so platforms/arm/samd51, no clue where they want that configuration header to be, don't know what the platform flags are, defining the parameters is not a bulletpoint, what's an extar, where's that file go, how do I know if I need other header files, this is not useful
I'm gonna say this qualifies as not having a guide for how to add support, because you would need to already know how to add support for this guide to mean anything, and retrospective instructions are junk
@karmic kite Real NYers express distance in time. It will take you an hour to get here. It will take you an hour to get there. I guess travel time on subway or buses and the delays are factored in where we round up everything to the nearest hour. Double or triple the time if you are driving, and only going a few miles. No one will be able to do the NYC Kessel run in less than 30 parsecs.
so you measure the kessel run in distance but standard speed in time?
Doesn't the UK use mpg for petrol?
although driving at 65 kilometers per hour at speed limit sounds faster to us
towns are usually 50-60 km/h highways 100-120km/h which is nice as 120/2 =60. going speed limit I know I can divide distance by 2 and get time
I drive 200 miles to get the kid at school. It can be 3 1/2 hours to 5 hours time, most of it with delays in city bounds.
daily?!?
no, she is at uni
yeah that could be frustrating
so another fun fact: most NYers or those from big cities with a mass transit system, do not drive and get a car until absolutely necessary...road test is tough and car insurance is expensive.
We just got rid of our public transit, so if you don't have a vehicle good luck getting between cities, or any town for that matter
our mass transit system is a constant fixer-upper. Mainly, upping the fares every few years.
public transit, seems like it would be so easy to optimize with a bunch of graph theory/simulation analysis. but in practise it never works well
Common sense is no where in the equation. Like even when trying to get a stop sign or traffic light put up at a dangerous intersection. The answer is install speed bumps in the road.
used to work at a minesite, which was a cluster of 4 mines. 630-730 am every day has about 5500 people trying to merge onto and off of a highway at an uncontrolled intersection. They've been trying to get an overpass for decades, but the DoT doesn't think the 1 hour a day is worth the cost
L/100km is a great unit from a usibility perspective
liters per hectokilometre?
deciMegametre
so many of these embeds offer no value and take up a huge amount of screen space
functionally the math works out the same but flipped from mpg or km/L, but because the number is presented in terms of a large unit of distance instead of a small one, it tends to convey a better understanding of what that number means
if you go 100 km you will burn this many liters
with mpg, well if you're getting 35mpg, and you go 105 miles, you'll burn 3 gallons, but what if you go 45
The really common unit here is per tank:P EG. My truck can go 800km/tank (60L)
and of course it's all junk because of how little reality the number represents
if you were going a certain speed and held that speed on level ground in a straight line for 100km, you'd get an amount of fuel that through a process that's in part unique to that car, relates to the efficiency rating
I was shopping at the pop-up shops in SoHo then wanted to go to Adorama and found myself on Varick street so I waved as I walked by
?
Future toffee
needs more proximity to me
I wonder what Adafruit plans to do with Show N Tell since Google is shutting down Google+ and from what I can tell is shutting down Google Hangouts
Nice Andon. My parents made toffee and peppermint bark yesterday. As well as chocolate covered pretzels and nuts. And caramel corn too!!!!
there's a lot of video chat services, it'll just get more complicated to run the show
why would they shut down hangouts?
it's not though
I wonder if theyre gonna push duo more as well
Google is shutting down several things with Google+
Including several of their chat clients/services.
as of december 2nd google's statements are that a bunch of confusing things will happen but 'hangouts' is not planned to go away
it does look like they're going to sunset the current 'hangouts' app on android, but it's functions will be replaced with hangouts chat and hangouts meet, btw hangouts meet would probably be a suitable replacement for adafruit as it's an enterprise grade product designed for corporate video conferenceing with the ability to allow outsiders into the confrence
especially if adafruit is using G-apps already, which I have no idea about
Ok
it's always hard to say what google will do, because they always give up on everything at some weird point
honestly G+ shutting down is one of the most explicable things google has ever said they're going to will have done
I've heard that G+ is a nightmare on the backend
I'm sure every social media platform is but I'll bet G+ is better than twitter and facebook in that regard
the thing with G+ is they've recently discovered a small number of not that bad insecurities, and they're just not interested in doing work to maintain the site. So far as I know they've silently fixed all the problems but this isn't going to keep happening
running any public facing service is a continuous process of discovering flaws and fixing them while at the same time trying to justify development to remain attractive, drive growth, and improve the incumbency value
G+ has failed, so it's hard to justify peoplehours to fix it, let alone improve it
that and as mentioned google gives up on everything weirdly, like when three years after everybody was exhausted about being mad about G+ integration on youtube, they suddenly pulled it
which was probably for the same reason they are pulling old style annotations, they want to change something, the new thing won't be compatible, and they can't justify doing the work to keep [G+|annotations|credits|googlechat|googletalk|googlevoice|etc] compatible with the thing that's changing
remember when google was an RSS provider?
remember Allo?
I wonder when they'll inexplicable abandon Google Home
remember Google Now? which is different enough from Google Assistant to not count; because Google Now was useful and Google Assistant can't do anything
It really comes down to a simple thing -- online services are awesome. They can really optimize and change the way we do business and interact with others.
But it is best to diversify. To have fallback options. If you find yourself becoming dependent on any one, it's too easy for a brief service interruption or a change in corporate "direction" to have unfortunate results.
I can't speak for Adafruit, but given their track record, I'm sure they have demo'ed other services and are prepared to switch gears if needed.
As with any change, there are growing pains and a learning curve. But we're engineers. We'll figure it out.
But wait, there's more! Fudge!
As far as Google Hangouts goes, that's just the easiest and most convenient. There are others.
I'm going to be making white chocolate/peppermint fudge next, and my mom's been making cookies.
My nerdy present to my wife came in the mail today. The keycap was ordered months ago!
White Chocolate Peppermint Fudge. Hope it turns out as well as it looks.
just remember that trycatches do weird things with stacks and heaps, that significantly impact performance
so i'm looking at this esp8266 board and i realize: at 160mhz it is actually faster then my first desktop pc lol
More ram too?
well my first pc had 48MB
and i ordered some esp8266 with extra ram
i think all things added up these might just beat that 48 lol
I wonder if you could like, run DOS, on an ESP
Mmmm If it could like to run NodeJS we could have proper web host.
dude! that would be awesome for retro gaming ๐
i imagine testing that would require in depth knowledge of the hardware though
lol at first glance, it seems less then 100% impossible ๐
unlikely
retro games are not just hard compiled for the platform, but they're also super sensitive to specs
there's a lot of old games that don't correctly account for CPU speed, like Space Invaders or Need For Speed, where if you run them on a system that is too fast, they run at a constant relationship to the clock speed not real time
lol yeah i'm familiar with those
if it can be done on a STM32 it can probably be done on a SAMD21
i got some smt32 on the way
the M0 and M3 tiers are extremely similar at a functional level, the M3 is mostly specced higher
lol at $2 a piece, i couldn't resist
Heh, my first computer had 768B of RAM and a 1MHz CPU. Slightly more RAM than an ATtiny85, but slower.
It had 7 memory sockets, 1 held a ROM the other 6 held 6810 128-byte static RAM chips.
lol
with neopixels, if you had one color set to 127, and you increased it to 128, would that be perceptible?
Applying for a job rn
Requirements:
Programming Languages
- SQL
- Assembler
- C++
- HTML
- Java Script
- JAVA
- Object-Oriented programming languages
some of these things are not like the others
Bahahahahaha!
I just remember having a convo once about using C# for program & kid wanted to use Java
@karmic kite you will work by google? Then be aware that they are steeling your ideas... ๐
almost every company I've worked for has a IP rights clause somewhere, however if I developed something on the side they've looked the other way
ok
also yes C# > Java
jumping around the censor isn't advisable. unless your talking about the programming language
@canadians is there a hobbiest level for IoT LTE/GSM, I've been looking at Telus and Rogers but it seems they only deal with Business
Any chance one'a yalls knows about super cheap chromebooks?
just looking at the ones Google has listed, there's 4 under 200$, one of which appears to be a previous model year of another so there's 3, a lenovo, an acer, and an HP
so the HP and Lenovo run a 1.6/2.48GHz celeron N3060 with Intel HD400
the acer runs an N3350 at roughly the same speeds, I'll have to find what iGPU is on that
HD500
allways connected pc's are cool
can't find any source of information that recognizes both the HD 400 and HD 500 with useful information like benchmarks
most can't even be bothered to have mostly complete specs for both
400 has a higher base clock, 500 has a higher turbo, I think that would mean the 500 is generally more efficient but can boost up just as high
N3350 has a lower base clock with a higher turbo, same so the newer processor should have better battery life for roughly the same performance
oh, N3350 has VT-d ( a specific kind of virtualization), that's nice
so it's a pretty marginal difference in the specs but there's no useful data on how that relates to actual performance
At least from looking at geekbench scores, the N3350 is a considerably faster chip.
@karmic kite BELL has some IOT stuff too but is also limited to business applications
Ah you got my hopes up.. I'm out of their coverage radius
Hello everyone, how are you guys/gals organizing your electronics in your lap/office ?!
looking at the totes beside me... very poorly ๐
My lab bench looks like a tornado came through and then a hurricane came through shortly afterwards.
Our local makerspace is getting a machine shop... thinking I'm gonna see if I can start making meccano parts, as they're impossible to find with out paying insane amounts for them
I'd be interested in hearing about it
machining sheet metal is really tricky so the way you solve a number of problems will be neat
I'm thinking superglue the sheet down to some MDF, spot & drill the holes, Vbit the profile, then cut the profile on a bandsaw, maybe
I'm thinking rapid press for the holes
I do have a cnc I could use as well
I seen a guy punching holes by hand which is an idea
if they have a CNC, it'd save time to have it spot drill the hole positions so you can punch them with less layout
A CNC with the structure to be a good press would be a bad CNC
And trying to do pressing with one that isnt can cause twisting that may permanently damage it
Consider the design of an arbor press https://images.homedepot-static.com/productImages/a883f76a-e5ad-4d55-8f35-100970a6ac67/svn/steel-core-metalworking-tools-43116-64_1000.jpg
This can be anywhere from 500lbs to a few tons
Maybe just a drill then
Sure, buy drilling through sheet metal tends to snag, grab the bit, autofeed, its not fun and thats why universal bits are preferable for sheet metal, or punching
Spot drilling can be handled safely by the CNC and eliminate the need to do complex layout by hand to get the holes in the right place
Or you could glue/clamp the sheet between wood
Btw, nibblers are real cool
https://youtu.be/kkH0AxF7soU skip the first 30 seconds or so
We take a look at the Hitachi CN18DSLP4 18V Lithium Ion cordless nibbler. Visit us on the net: http://www.toolsinaction.com Watch us live on Periscope http:/...
The way nimblers work is inside there, there's a small metal rod going up and down, and it has a section that is a smaller radius with sharp transitions
It works basically like a paper holepunch, but fast
They're not great at straight lines, but they're super great at freehand
Those are fun to use when they're in good shape, and it can be amusing to watch them machine gun out a bunch of small crescents of sheet metal
The one we have at work just loves to bind up, and they're endlessly frustrating when they need servicing :\
All tools need to be cared for, cutting tools most of all
Absolutely. I should probably get one of the mechanical guys to do that ๐
The nibbler part is probably a consumable, and maybe the plate it nibbles against
lots of lubricant I'm thinking would be helpful
Probably not for jamming
for longevity though
The nibbler rod and plate work as teeth, so they wear down, and once they wear down there's enough play to jam instead of cutting
It might be possible to sharpen the plate, but the rod is usually too small
I may also just see what tools we end up getting in the shop and use whats available to me
sure, but if they haven't spent the whole budget, now may be exactly when they need somebody to have an opinion
if you don't already know what goes into a shop, it can be hard to know what's actually useful
belt sanders and disk sanders are extremely useful, and belt grinders are super great
@old fjord https://www.aliexpress.com/item/Mini-Stainless-Steel-Foot-Pedal-Switch-Controller-Tattoo-Power-Supply-Machine-Footswitch-5ft-Clip-Cord-Tattoo/32845006949.html?
this must be the most obvious makeshift contraption I've yet seen on ali
Smarter Shopping, Better Living! Aliexpress.com
but it's also geniusly simple
it's two metal plates mounted with rubber bushings on two bolts, each plate has continuity with one plate, and the wire is connected with eyelets, on one side under the bushing in contact with the top plate, on the other side above the bushing in contact with the bolt, you step on the top plate to make it contact the bottom plate
I think that is a single throw switch. ๐
Similarly...
Ugh. Sometimes when things just dont work out and it really feels to pull all the hair out and cry. ๐ ๐
I know that feel.
๐ช
I have been trying to run Node and use NPM and Yarn on an Android Device, I can't just install the Packages.
I left the JS world behind years ago and never looked back.
I looked at it a while back, didn't really interest me.
What do you use then for backend ?
Python, JSP, Java, various hand-rolled things. I've built a lot of backends over the years.
Mmmm I see. I have heard of Python and Java being used. But Mmm JSP Its a new name. ๐ค
JSP is also Java, but with a well-defined set of interfaces so it can be automatically deployed by special web servers known as "app servers".
Back in the old days, I used plain old C, as I needed the performance to support large numbers of clients with the available hardware (a Sun 4/100 in those days ran at 14.28MHz โ slower than an Arduino). Some people then went to Perl (which I didn't really like) or ColdFusion, but my clients couldn't afford that, so I stuck with C. I briefly went to PHP (which basically replaced ColdFusion) along with some frameworks such as Django, but I wasn't really a fan of that either. I dabbled with Ruby as well, but never built any production code with it.
I just read that Discord is only charging developers 10% to host video games through their store vs 30% for STEAM (valve softeware) so if you are in game dev you might want to check them out
but at what cost, market oppritunity
I for one would never install a game from discord, and I have discord, there will be others like me, and then there's the massive population who don't have discord. Steam has been building their customerbase for 15 years, tens if not hundreds of millions of people have steam ready to purchase your game
as of may discord claims to have 19 million active users with a peak simultaneous load of 8 million
they also claim to have over 130 million registered users, so it's safe to say their numbers are inflated or 85% of users never come back
discord may give you more of the revenue, currently, but you'll sell a fraction as much and they can change the deal at any time
getting games on the platform is more valuable than profits, they need to bootstrap it, so of course it'd cost less now, and of course it will cost the same later
epic and gog are also attracting attention as well lately
never heard of epic, but gog gets by via slant non-competition. The bulk of their library is stuff they updated
epic == fortnite
epic made money from making fortnite and just announced their own store not long ago
Epic Games have a launcher what Unreal engine developers use
there is unreal studio
what is now 4.21
8 days ago ๐
this is dumb
why do you say this?
Offer constructive comments
also subnautica is free right now apparently
I recommend waiting until it's not and paying them for a great game
if that's hard, take it for free
it give options and customers to choose
game devs can still release games to other stores too
I'd be playing subnautica now if my hands weren't in pain
cross play between platforms was announced too from epic games
Just a general reference -- I found this PDF guide on how to use digital calipers effectively, figured I'd share it here: https://littlemachineshop.com/images/gallery/instructions/usingcalipers.pdf
I've always wondered how close digital calipers are, compared to the accuracy that they say they are
these are good usage tips, here's another
Every tool has a job and in measurement devices that job is precision, accuracy, and stability.
Precision is how many digits down it goes, different tools are designed for different precision.
Accuracy is how repeatable the measurements are, a tool with high accuracy will measure the same thing hundreds of times and give the same result within the scope of it's precision.
A very stable tool is able to maintain a consistent calibration for long periods.
A good set of gage blocks treated well can maintain 0.025 thou or better accuracy over decades, a good micrometer can maintain 0.1 thou accuracy as long as the temperature doesn't change significantly.
A very good caliper should only be trusted to an accuracy of maybe a thou for a day. Calipers have the advantage of being much faster and easier to use.
Understand the how the dimensions of your project matter, and how the capacity of your tools influence what you think they are
and if you have cheap measuring instruments, don't expect too much from what they tell you
last tip: until you're a physicist, all measurements are relative, so if you can't compare a tool with something highly reputible, the tool's reading can't be highly reputible. If you care about precisions on the order of single thou and below, buy a gage block, I recommend something like 1 inch but the number literally doesn't matter, from a reputible brand and seller
this is highly excessive, but what ever your calipers say, this block is one inch long https://www.grainger.com/product/MITUTOYO-Steel-1ARU5
this may be the most niche question ever asked
do any of you remember what the plot of Steel Battalion was?
Not a clue, but the controller looks cool
if you're still on windows XP it's a very good controller, all original xbox controllers are native USB devices as long as someone has put out a driver and there was an extremely good driver that never got usefully updated to 7 before the midsection of my controller was stolen from my trunk
the wing sections are just switches connected by a big IDC connector, so without the midsection it was rendered useless
Hmmm, it might be able to be rebuilt.
Well, more accurately. It can be rebuilt, it's just a matter of how difficult it would be
mine couldn't, that was years ago and it's all gone now, but if you had one it probably wouldn't be hard to replace the control board to make it more generic, it is entirely just a bunch of switches and one rotary switch
not sure of a good way to replicate the jewel switch faces, probably the best way would be to 3D print a master, pretty it up real well, then make a mold and cast dyed... lost a word... resin
the original jewels were of course hollow injection molded plastic, it'd be tough to make hollow replacements but probably not necessary either. Just making the controller from scratch would come down to replicating the joystick mechanism, which I can tell you was not quality stuff
the one on the left is single axis, which is perfectly fine, but it's just a potentiometer with a yoke mechanism https://i.imgur.com/oUQvggl.jpg
and the one on the right is the same with a double yoke
it'd be real easy to build a better joystick
I'm sure you can get some sort of "jewel" switch face from somewhere. They're moderately common for toys and stuff
definitely, I was thinking more in the terms of replicating the specific aesthetic of the original controller, jewel faces are the kind of thing where if you buy it, you abandon a lot of choice in what you end up with
No lie there
I do want to build a custom controller with toggle switches and buttons and all that sort of thing. But, annoyingly, toggle switches aren't something game controllers come with, and most games don't have anything to support it
yea they're much more event driven
in the XP driver it had modal stuff in the driver, so you could configure what button 27 did when switch 5 was ON vs OFF
I had a setup where two switches would change the shifter between cruise control, 5 speed manual, and off, and then in 5 speed manual mode the position of the shifter ranged the throttle pedal in 20% increments
so I could put it in manual, drop it into first, and have the entire pedal throw to control 0~19% throttle
Wouldn't be terribly difficult to do in the controller for an Arduino
Heh, a quick peek found a library that'll make a 32u4 appear as one, two, or three game controllers
yep
https://www.digikey.com/en/articles/techzone/2017/sep/design-an-accurate-miniature-joystick
Not sure how much of this would apply to a joystick that rotates, this appears to be based on those sliding nub joysticks; and if I was making a gaming device I'd probably want to make the MCU responsible for monitoring the sensors directly just to eliminate communication time between the MCU and the joystick host
A single-chip Hall effect sensor and a magnetic assembly can form a miniature 2-axis joystick with high sensitivity, smooth operation, low power, and low cost.
I figure with the right sensor chip you could have a 2 yoke joystick axis isolator system with a magnet on each and the sensor would trck the magnetic field perpindicular to the axis
For any techies, Is a Quadro K620 worth it for 80$?
https://gpu.userbenchmark.com/Compare/Nvidia-Quadro-K620-vs-Intel-UHD-Graphics-630-Desktop-Coffee-Lake-i5-i7/2840vsm356797
no, not even a little bit
(~60$ USD)
nVidia nomenclature rules, the last two digits are where it fits within the generation, the preceeding digits are the generation. Anything under a #60 is garbage put out to exploit people who don't understand graphics cards
60s vary from generation to generation, I think the 660 was decent but I'm pretty sure they're out of production, the 760 was pretty mediocre but the 770 (what I have) is definitely still relevant if you can find one. There mostly weren't 800s except for a couple of mislabelled mobile devices that were I think actually in the very low end of the 900 generation. the 960 through 980 are great
the quadro series doesn't have any above 20 in the last two digits...
you can ignore the words
for one, that's not actually a quadro
quadro are professional grade cards and they start around 1200$
keep in mind that the last two digits being 20 means "this product exists to exploit uninformed consumers"
also Ti is basically V2, they're a little bit more optimal, clocked a little higher, sometimes quite a lot better but overall the same grade of thing
I am looking at professional cards ๐ This one is on our local classifieds pulled out of a workstation from a business that went under, 2014 model. looks about 3-400$ on ebay
half the performance of a 660, a third the performance of a 770
I wouldn't pay 15$ for it
it does manage to be marginally better than a 1030, which is exactly the same kind of exploitation product
the key fact remains that the 620 is entirely without merit unless you already own it and don't have any other solution for getting a video feed out of the computer
current is the 10 series, the last two digits represent basically price brackets
80 is the highest end card within a generation
although in the 600 and 700 generation the 80 cards weren't particularly good
I've been debating on a Quadro P5000 or 1080TI for a while, Some of the graphics software I use only supports quadro cards
And was thinking something to hold me over for a bit, as I'm still running HD630 ๐
Once in a while, you can grab a 1080 for cheap deal
I'm using my 650 until I can update mine
My mobo graphics is about as good as a 620, I believe
It's a newer mobo than graphics
intel graphics are pretty decent for what they are
Its kinda why I left this part out of the build... but been pushing it lately. Especially with tensorflow
I'm not going to suggest that any listing on ebay is trustworthy, do your due diligence, but 770s seem to be available at that price
according to passmark the 620 is twice as good as the HD630, and the 770 is 3 times better than the 620
at 200$ I can almost get a 1050TI
right but that link is 70$
+120$ shipping
7.68$ shipping
there will be other listings
there are multiple in the price range you're looking at
Tariffs are a thing unfortunately. Although I am going near the boarder next week, I wonder if I buy it and bring it back if they'll make me pay a tariff for bringing it back
if you're coming from canada and will be crossing the border near Buffalo, Detroit, or I think American Vancouver you should be near a Microcenter, they carry a lot of graphics cards and are pretty good at not being a ripoff. Mine has started carrying a large selection of clearance open box super suspect items that are almost certainly burned returns from cryptominers, but they're cheap. At least in detroit most retail stores are geared to accept Canadian currency so pay cash and don't tell the border guard
if you're lucky the manufacturer's warrantee is still good on the open box stuff, but you'd probably need to pretend you're in america to use it
I wish I lived near a MicroCenter. The closest one for me is 4 and a half hours away
@jaunty jetty back to erlang the difference between ; and ,
';' means 'and'
',' means 'or'
ok i gotta leave this server
sup dude, how's it going
something happen?
well, I'm going to encourage you to chill, watch some cat videos
no, i cannot be on this server anymore
if u want to say something to me ever again, pm me now
i simply cannot contribute to what i consider to be poison
goodbye
my cat wrote the song, I just played the music and shot the video. music is available here: http://cyriak.co.uk/music.html
well, mmkay, can't say I don't understand the position
His profilepicture is of a skull, so I reckon he is the poison ๐ค
Very evil like
is it just me or does the windows iso download page (https://www.microsoft.com/en-us/software-download/windows10ISO) always takes forever to load?
It takes a while
i always wondered- is there a reason why?
Ugh I have been super confused with my Web Blogging apps working.
I have been constantly redesigning it. Oooof
Might sketch up a Proper Working Plan tomorrow
boooooooo
Reason #5762 to dislike amazon
To be fair, I see where they're coming from.
Aside from the obvious "advertising their own contracted professional services"....some jobs honestly should NOT be DIY'ed by someone who has no idea how to do them.
Plumbing. Electrical. Furnace/HVAC/gas lines. You might put yourself into a very dangerous situation at the worst, at the least you're not up to code.
Sure, you might shrug the not-code-compliant off, but your insurance company won't if you make a claim.
For many of these, if you DIY correctly (pull a permit and get an inspection), your insurance will cover it. Personally, I do most of my electrical and plumbing myself, and some HVAC, and hand the rest (and all gas work) off to contractors.
Ah, but you have an idea how to do it.
Taught teenagers plumbing, electrical, and tile work at the local teen center a few years ago. The inspector said he was going to be extra critical, since peoples' kids were there. He pulled the cover off one of the outlet boxes and said "this was obviously not done by professionals." The kids' faces fell. Then he continued "this was done by people who cared about what they were doing. All the wires come in neatly, are the right length, and everything is well arranged. I don't need to see any more. You pass." They were thrilled.
For every one of you, there are 100 weekend warriors who get in way over their head. Or worse, they never realized they were out of their depth.
When it comes to a $30 microcontroller, we call that a learning experience. When it comes to a $300k home + the lives within, that's....far more serious.
True enough, but I wouldn't recommend amazon to find contractors (or anything else, for that matter).
Oh golly no.
Finding quality contractors is an art in and of itself. I'll generally start with local recommendations and try a few until I happen across a really good one, then get referrals from them. Yup, I'll ask a good plumber for a good electrician or painter or whatever: it seems to me the good ones seem to know each other.
I can agree with the "some jobs should be left to professionals" position, but the structure of that add is DON'T DIY (pay us instead), and what most people are going to take from that isn't that amazon will let you rent a plumber but that DIYing is bad
most people aren't even going to notice that the word 'home' in pastel blue on a white background between the bold black "amazon" and the bold black "services" exists, so they're not going to see the context of this being in relation to home repair or modification, just DIY in the abstract
Yup, well put, and I agree.
if we're lucky most people are cognizant enough of Amazon's existence as a giant emotionless megacorp that desires only to recieve money from them to not care what Amazon thinks about DIY in any context, but people are impressionable and easy to manipulate, not paying attention. X in media doesn't cause people to X, but frequency of X in media does change how people feel about X
also let me just say, Amazon, don't put pastel text on white background, some people don't see great and we like to read things too.
I read an article a little while back that it's not "Amazon" that chooses bizarre colors/layouts, that Bezos himself dictates exactly how the web site is to look.
I wish I could get paid >$3000 per second to make bad decisions all day
that could provide a 30k$ salary to over 3 million people
He's certainly not worth that!
nearly all of his wealth exists as worth, if you liquidated his holdings and properties you'd get a dollar value but he's probably not that liquid so he may only have a few hundred million dollars available on the short timeline of weeks
I was being facetious and considering his worth as a human being.
sure but, then we're being metafacetious, because no human should ever be considered worth that
Quite.
it's like, if they legitimately were, the fact of it would be proof of their undeserving, because such a state must result from some kind of manipulation and abuse of trust
what is the policy on discussing non-arduino based projects
we do so frequently with wreckless abandon
great
cool! a radiation tolerant M0 -- looks like M0's are headed to space ๐ https://www.cobhamaes.com/pagesproduct/prods-hirel-arm.cfm https://www.cobham.com/advanced-electronic-solutions/news/cobham-announces-qml-qqplus-qualification-of-their-radtolerant-microcontroller/
Aeroflex Colorado Springs Circuit Card Assembly capabilities consist of full assembly, test and coat in a high mix/low to medium volume operation
these may be a bit more expensive than an ATMEL SAMD21 ๐
Truly off-topic, can anyone recommend a Discord channel for other kinds of Making? I am interested in getting into epoxy projects and I'm wondering if there's a good Discord that covers that kind of stuff
i'm going to have to steal that meme
Only other server I has is Waze & local maker's guild
don't you hate it when you go to microcenter to buy some female headers and literally the only category of female headers they have is sets of arduino uno headers with passthrough length pins, one set for 3$
I could get 50 pieces of 16 pin female headers for 7.29$ on amazon, or 10 pieces of 40 pin transboard low profile passthroughs for 6.49$
@jaunty jetty It helps to shop around.
the 50 pieces of 16 pin is actually one of the better deals I've seen
it's a bit less than half the per pin cost for that volume at digikey
unfortunately microcenter is the only game in town for brick & mortar
I think I bought my last set of headers from Tindie, where they were cheap in available in 6 different colors.
@jaunty jetty I....watch my words carefully when talking about Microcenter. I appreciate having a B&M store that carries low-margin parts.
But my Microcenter, man...I've been going there a few times a year since 2010. Every time, the store looks like a tornado hit it. The clerks can't find half the stuff I ask for. They gestured towards the small parts bins and said "ha. good luck finding it."
Online orders? hmm. I tried that. A few years ago, I needed an adafruit part ASAP. I placed orders for a NJ microcenter and three different NY microcenters (two in NYC, one just north of the city).
Every order was cancelled because they didn't have the part -- although the online part inventory stated a plentiful supply. I simply ordered from Adafruit or Digikey, and ate the shipping + delay.
oh yea, the hobby electronics section is literally just about 30 sections of pegs with items loosely segregated by shifting categories
currently there'sa 15 foot section of "most'a the stuff", a 4ft of heatsrhink, a 4ft of motory bits, a 4ft of connectory bits, and then the rest of it
oh and an endcap of exclusively adafruit products, that contains about 30% of the adafruit products they have
Since then, I now buy 1-3x spares of every "necessary" part just so I'm never caught high and dry. Of course, depends on the item and the cost.
Digikey + Adafruit for sensitive parts, Amazon + ebay for the rest I can wait longer for.
it'd be nice if the microcenter corporation hired like two people in the whole of north america who knew stuff about stuff and could try to make their selection more sensible
My poor USPS mail carrier is so dumbfounded when I have a constant stream of packages from china coming....I forget a lot of what I ordered, since it was 2-3 months ago.
they've got at least 20 SKUs of male headers including a lot of different bulks and specialty stuff, but 3 female header SKUs that are all uno headers
And here's the thing -- I want to speak better & support Microcenter. Seriously. But after getting burned so many times, I gave up. Might as well realize that any issue I have has a 2+ day resolution time.
eh, you don't need to want that
and I wouldn't buy anything from their site
it's a physical store, the whole reason to do business with them is that it's physically there, you can just go make a bad decision quickly
if I could have bought a small pack of female headers for like 5$, even if it was a ripoff, I would have, but the selection was just too poor for my uses
I feel ya.
also all their PC cases have tempered glass sidepanels, and I hate that
anyway, it's like mcdonalds
you don't go to mcdonalds because it's good, it kinda is but it's not that good, and you don't go because it's cheap, it looks like it's cheap but it's super not
but it is there
and if you wanted a better tasting cost effective burger, you'd have to put in more work
I had a radio shack within walking distance before they closed.
Every time I walked in there, it was high pressure cell phone sales. "No thanks, I already have one." Them: "Need a second one?" Me: "All I need is -electronic part name here-" Them: ---blank stare--- "Um, stuff is in the back, I guess."
But I felt like that DIYer from Parks & Recreation going into a home improvement store. "No, sir, you can't help me. I know more than you."
My local Radio Snack was better than that. I bought out their carded stock when they closed, now I have my own Radio Snack.
@late fulcrum I did that many a time. Buy one get one free, 90% off.
My receipts were 20-30ft long
I have a box full of switches I think were so cheap, I feel almost guilty. almost
that's getting close to a whole roll of thermal paper, you should build a time machine and try to hit that
My only regret is that I didn't just clean them out. I was judicious, bought what I thought I could reasonably use as backups and spares.
....but mostly cleaned them out on those things
What it's like to have a Radio Snack inside your house
They threw in the pegs and floppies (plastic pieces that hold the tags), so all I had to do was hang some pegboard.
are these real?
for certain definitions of real yes
all LCD screens use a system of polarizers to achieve light exclusion through the LCD panel, there's a back light, then a polarizer, then the LCD which is a variable polarizer, then a depolarizer* that reveals the color. Because of this, if you view a screen through polarized sunglasses it all goes a bit wonky as the sunglasses block out parts that are polarized different to themselves
if you intentionally created a pair of glasses with the polarization that conflicts with common television polarizer setups, compatible televisions would, while you wear the sunglasses right side up and hold your head matching the orientation of the television, see some dimming effect based on the quality of the glasses' polarization, the quality of the TV's polarization, and viewing angle
significant deviation from being correctly on plane and in front of and aligned with the TV will reduce the effect
actual LED, plasma, CRT, OLED, and LCD TVs where the polarizers are oriented differently than expected will not be blocked correctly
thank you for that info
LED backlit LCDs will be effected, so will differing CCFL LCDs, and I think plasmas, but direct LED and OLD do not use an LCD panel to cut color from a backlight. Also most phones even if they have a backlight system with polarizers, tend to have much more complicated systems and make use of spiral polarization, so they do not uniformly interfere with sunglasses but you might see discoloration
is the opposite possible? a screen that cant be seen without glasses filtering it?
sounds amazing
Do you have old displays kicking around? Want to make sure your private windows are truly private? Check this out... Mack Weldon: Visit http://bit.ly/1NcoxLH...
OMG
you will likely see a reduction in color quality and this shouldn't be considered any degree of security
btw @idle iron this is essentially how passive and cinema stereoscopic works, in the cinema they have two projectors projecting polarized images over the whole screen, and the glasses have different filters that only let one through, there's no loss in fidelity and if configured right there's no loss in brightness. For passive home 3D the rows or columns of pixels are alternately polarized and you have the same kind of glasses matched for the display. Because it's spliting the screen area between two images you lose half the vertical or horizontal resolution and the two images can't exist in the same position so you're relying on the visual cortex to stitch over the missing space. It's a bad experience all around and each eye gets half as much light (less actually) so the screen has to be super bright to look kinda dim.
Active home 3D uses time spliting instead, every other frame is sent to each eye, and the glasses use an LCD panel in each eye to alternately block it out in sync with the screen. This still causes dimming, and you lose half the frame rate. A good TV should be able to handle 60FPS, most don't, and a well mastered movie should be 24FPS, most home equipment can't even do that so a lot of movies aren't mastered right. In the end it's very common for badly mastered badly made 3D blurays to give the worst of all worlds packed together, and that's why everybody hates them.
sudden urge to replace all the springs in my mic boom with pneumatic pistons and cable linkages.....
not controlled, the pistons would be use some kind of spring valve and use vacuum pressure to resist being extended until sufficient force is applied to overcome the valve tension to alter their balance
Just a "maintain position" mechanism? I'd probably use hydraulics for that. Either way, you'd have to be careful about flow noise while repositioning.
hydraulics would be stronger but a lot more rigid, this is holding up a microphone so if it can absorb some vibration that's a good thing
what are the people... at the stock exchange...... actually doing?
like if we ignore all the future we've got now and go back to 1983, what are any of these people doing https://photos.smugmug.com/THE-AMERICAN-STOCK-EXCHANGE/THE-AMERICAN-STOCK-1/i-nPVJxL6/0/129f3b37/L/TF%2310-1983-L.jpg
They're pretending to be big wheels playing with other peoples' money. They're carefully watched. But the big money (and lack of oversight) is in the derivitives markets.
Interestingly, I was in NYC a little while ago, and saw they had lit the stock exchange building with red and green holiday lighting and went for a better look, whereupon I tripped over a little bit of history in the making: they were relocating the Fearless Girl statue quietly in the middle of the night.
oh was the lawsuit settled?
reading the wikipage, I'm always surprized by how insignificant change.org petitions are
A petition on Change.org asking for the statue to be made permanent gathered 2,500 signatures in its first 48 hours.
they're always brought up in a statement of "look how many people want this thing", and then are followed by the most irrelevant possible number of people
there are kickstarters that raise millions of dollars in the first 48 hours
I went out the next morning to see it in better light (and when photography was allowed) and actually met Harper, the little girl who was the impetus behind moving it.
these petitions rarely get over 80k people
according to wikipedia 80k people is 0.0247% of the united states, and these petitions along with being inactionable most of the time, tend to be global scope
until 5000BC, it was nearly 2% of all humans
if anything, every change.org petition has been proof that the petition should not be acted on
I wouldn't agree: many of the ones they've floated have been worthwhile, practical, and for the common good.
that's not what I'm talking about
the turnout on these petitions is so grossly negligable that if they were given consideration, the logical take is: people don't support this
That's more sampling error than a "mandate of the people", as I'm guessing you're well aware.
the entire petition falls within sampling error
the good side is that nobody values change.org petitions, they get no consideration because we know they're not representitive of anything
but they sure get brought up a lot, relative to their meaning
the petition in question got just over 40k signitures
0.46% of the population of New York City; and not limited to residents of NYC or the USA; so globally they got half a percent of the city's population. Boy people sure don't support that
I talked to a congressional staffer once and asked about the impact of petitions. She said that a petition (regardless of the number of signatures) was typically counted as "1 letter".
sounds about right
I've heard they also value actual letters higher than emails, and hand written letters higher than printed letters
there is I think merit in considering effort, to an extent, in bulk
That echoes what the staffer told me. Phone calls were between emails and letters.
were it me, knowing how I feel about phone calls, I'd put them between hand written letters and printed letters
I have a similar antipathy toward phone calls, but apparently congress takes the "Santa Claus" viewpoint that a letter is a real, physical object, and therefore counts more.
yea
backsplaining it, they can use a physical object ways that even an audio file or, dare I suggest, a cassette, are just impractical. In the end the goal of communicating with politicians is A, to inform them of your desires so they may seek to fulfill them as appropriate, and B, to arm them for achieving your desires
this is too close to politics
She said that for big issues, the letters are rapidly sorted into "pro", "con", and "other" piles, then the "pro" and "con" piles are simply weighed.
After that, I would put my letters to congress on heavyweight paper. ๐
I like the image that in every elected official's office there's a giant set of justice scales that they're constantly swapping piles of paper on and off
congresspeople get two scales so they can process the piles faster
My dad had an old broken GPS direction thing and asked if I could fix it. I told him even if it was fixed it is antiquated. Why would he need it if he has a working cell phone....
Having a device that works without requiring cell service is nice in certain portions of the world...
I live near the Canadian border. Sometimes my wife and I run up into Vancouver for a day trip. It's too much of an expense to get cell/data service turned on, but our GPS is enough to get us to/from our destinations.
One use case, I'm sure there are others
That is true. Thanks.
Also here in WA and a growing number of states, there are distracted driving rules and you're not allowed to be touching your cell phone while on the road.
Even if you're sitting at a stop sign/light.
i found this https://www.smartertravel.com/use-phone-gps-without-data/ havent read though it yet but looks like his phone will do GPS without cell data. he has iphone 8
augmented GPS transponders like your cellphone have vastly better performance than strictly GPS devices, especially after adding inertial tracking and coordinated location services
establishing a GPS lock is actually quite time consuming and energy intensive, so your phone isn't really using GPS that much. What it does instead is track cell towers, inertia, wifi networks, and look that up against a database of previous GPS locks to determine your approximate position accurately based on pasted data. Even just having that approximate location then makes establishing the GPS lock much faster because it already knows roughly where you are and just needs the GPS to fill in the fine detail rather than produce a location from whole cloth
as well it doesn't need those GPS locks frequently since it can use all those other factors to approximate your change in position from the last fix. Back in android 2.2 when they didn't have all the aGPS techniques down, and hadn't put together the coordinated location services yet, navigating on GPS did require a constant lock and would run down the largest battery in an hour
if you need better positioning than that you'll have to use RTK, https://en.wikipedia.org/wiki/Real-time_kinematic
you'll have to set up your own station, it'll only work in a certain range of the station, but you can get centimeter accuracy or better
oo i like that, would be cool for a robot
a robot like mabe your grain harvester
or a bot that moves around the green house auto watering plants
might be easier to put that on rails, or wires
I'm not sure if I trust signals to rural enough for that yet
Or are we talking about super powerful antenna GPS system?
I'm not sure which what you're talking about
like afraid of cancer from electromagnetic radiation?
Aside from health concerns
I'm talking about the precision one needs to remove operator from seat in harvestor
Seeing current of how wonky GPS gets rural, I'm wondering how that'll be resolved.
they use RTK
the farm has an RTK station on it, the harvester augments it's GPS position with the RTK, and this is how it's presently done
i think a good triangulation system would work too
Nevermind, I just caught about station tower
Most use triangulation, just albiet towers or satelite
So, just need power generator, a station, then all set!
I used to have this toy that was an RC cable car and a bunch of clamp on kinda hotwheels track sections but instead of a track the cable car rode in, the clap bits had a flat and then you attached plastic wire like for a weed wacker between the clapy bits, and the car would race along the wire however you had it set up
that looks like fun
i like that idea, that would be cool for the green house watering bot
yea so the clampy bits were like this
so make a heavy duty version of this and rig it up around your greenhouse for the watering bot to roam around on
I just read back to what started this convo, interesting article. Anyone else here plays around Waze or Google Maps?
I tried waze a couple weeks back, the UI was a mess and I had to get rid of it
i use google maps sometimes, never heard of waze
every time I panned the map it would randomly snap to show the most advertisements
there's never any excuse for having a map snap at all
? 1500s
Yea, the app UI sucks
i think they had adds in like 1800s also?
I also wasn't allowed to select my destinations because they weren't sufficiently valid, and everything did random things every time I did anything
Couldn't select destination? That's new one
anything Waze does google maps cant? is it just an alternative?
Waze (property of Google) is google maps times social media
Huh?