« first day (2089 days earlier)      last day (2755 days later) » 

12:00 AM
@ConorO'Brien Why
 
that this has to be specified is saddening
 
> temporary opening for an expert level computer programmer
Because expert level programmers just love to be hired temporarily
 
@ConorO'Brien Why is it saddening, web applications are a really common use for Python
 
@ASCII-only really
 
Yeah, Python/Django is like JS/Node and Ruby/Rails, all pretty popular frameworks
Maybe not node, something more like express/sails/yeoman
 
12:06 AM
@ASCII-only Flask is better :P
 
@Maltysen I haven't used Flask, so objectively, Django is better. :P
 
I haven't used either, so JavaScript is better
 
12:24 AM
QPixel is the best website, so Rails is better
 
true
I'm going to post this if there are no pressing matters
 
speaking of which I had a truly marvelous idea for a web framework using Pytek's upcoming C and JS transpiling
 
@quartata ...but this message box is too small to contain it?
 
yes
 
@ConorO'Brien I would suggest rewording "Given a list of non-negative numbers of length N," to "Given a list of N non-negative numbers,".
 
12:31 AM
ah, nice catch
 
CMC: write a CMC in only semantic primes
 
@ConorO'Brien what happens if its like [123]
do I pad it down to 1
or what
 
@Maltysen "You may assume that N is greater than or equal to the number of digits of the largest number in the list."
 
derp
 
12:33 AM
can I take the numbers as strings?
like ["1", "2", "3"]
 
idk if I should
why not
 
@ConorO'Brien I missed that on the first read as well. Maybe there's another/better way to emphasize that.
 
I bolded it
 
@ConorO'Brien wait can I do like this: [' 4', ' 5', ' 6']
 
@Maltysen hahahahaha no.
 
12:46 AM
ok
 
Anonymous
@El'endiaStarman You should try Flask. It'll change your life. Maybe. Or maybe you'll just think it's neat.
 
Flask isn't very scalable or so I thought
 
@quartata the dev server isn't
but you can run it on gunicorn, appengine, heroku, etc.
 
Eh. I might check it out someday, but realistically, I already know one Python-based web framework, I don't need to learn another to do the same stuff. :P
At least, not at present.
 
Anonymous
Yeah Flask isn't meant to be a scalable web solution. It's meant to work in tandem with scalable solutions to actually create and serve the content.
 
4+4+4+4+4...=e
 
feedback?
 
it looks like V8L
 
@Downgoat looks like V0L
what is it?
 
VSL...
pls halp text logo is hard ;_;
 
1:39 AM
or V upside-down sigma
 
what about:
 
It definitely needs some space around the bottom-right and upper-left bends of the S.
 
letter spacing is your friend
 
i suck at this .___.
 
@Downgoat it just looks like you messed up now
 
1:40 AM
i think the merging idea just doesn't work when s is the middle letter
 
also the un-inverted colors was better
 
@Maltysen first one?
@xnor ok, true
 
@Downgoat yeah
 
why are you making a logo when you haven't started the thing
 
what is VSL?
 
1:42 AM
also make the s smaller
 
@quartata I tried pointing this out multiple times.
 
@quartata the logo's the funnest part
 
@PhiNotPi no naming it is :P
 
;_; why can we not just name it after noun
its much easier this way
idea: name language triangle
imagine all the beautiful triangle logos we could make
 
@Downgoat what is VSL
@Downgoat like the illluminati?
 
1:44 AM
So far it seems to be an odd logo.
 
Anonymous
I still don't know what VSL is
 
let's start guessing
 
Anonymous
Very Small Languages
 
Very Slow Language
 
it's his new scripting language
 
1:45 AM
Vanilla Strawberry Lollipops
 
Virtual Software Library
 
Vividly Surreal Landscapes
 
Vertical Stupid Lynches
 
Vitsy's Smaller Lapdog
 
Viscous Shock Layer
 
1:45 AM
Vector Sum Lengths
 
Vampire Stake Lathe
 
Very Short List
 
Anonymous
Violent Sociopath Lemur
 
Vulcan Study of Logic
 
@Mego Looked up Flask...its a bit too simplistic for me. I like my routes.py and models.py
sorry for breaking the chain :)
 
Anonymous
1:46 AM
@NathanMerrill But @app.route! Routing is done via decorators!
 
Anonymous
Isn't that cool?
 
and third parties can make their own routings
 
^ this is why I don't want it
 
like @socketio.on
django doesn't even have decent socketio integration
 
Anonymous
Also you get SQLAlchemy for DB access
 
1:48 AM
Wow that's a lot of VSL backronyms.
 
or I get models.py, which builds a database for me
 
Anonymous
Or you use literally any Python DB library you want with Flask
 
pls halp now it look like shit
or does it idk
 
well at least I can read it now
 
Anonymous
@Downgoat But what does VSL stand for?
 
1:49 AM
About the backronyms: Small yes, slow hopefully not
 
versatile scripting language
 
Anonymous
But why male models?
 
@Mego You can use any one of several DB backends and Django does the work for you.
 
0/10 red also we should rename it first
 
1:50 AM
;_; y u haet red
 
@El'endiaStarman you don't really get the choice of your own DB library (unless you want to totally break out of the Django framework)
 
I am red
 
Anonymous
@El'endiaStarman That's not a good thing. I want my libraries to do exactly one thing, so I can mix-and-match them as I please. Cohesion is important.
 
by transitive law of equality you haet me ;_;
oh shit i did a zyabin :P
 
Because red usually means bad
 
1:51 AM
Q____Q
 
@Downgoat how about you mix up the gradients?
or add some other cool effects
 
??
 
@Downgoat Are you asking for a suspension?! (I kid.)
 
@Mego I'm not sure I follow. Are you against frameworks in general?
 
@NathanMerrill Oh, I see the point.
 
1:52 AM
I mean, most frameworks provide more than one thing
 
to be fair, however bad my logo is, this was the previous contender:
 
Technically, you still can use other DB libraries with Django, I think.
 
We need some fancy name and catchphrase like "Edam: The language that was made backwards"
 
@Downgoat 10/10 very readable
 
Anonymous
@NathanMerrill I am against non-modular frameworks. If you can't replace parts of the framework at will, then it's not very useful.
 
1:53 AM
you realize that modular means non-cohesive?
that's the tradeoff
with django everything just works great together
 
I prefer the nginx approach to modularism (in contrast to apache). This applies to every software for me.
 
@NathanMerrill flask works pretty great too
 
but if I have different parts doing different things, you either have to find or write the glue between the parts
 
@ASCII-only Scriptgoat: bleeeeet bleeeeeeeeeeeet;
 
"Everything just works" is rarely true, and always just to a point.
 
Anonymous
1:54 AM
@NathanMerrill No, it doesn't. Each of the modules does exactly one thing, and the framework acts as the connector between the modules. A framework should do nothing more than facilitate communication between independent modules.
 
@NathanMerrill but the third party people (like flask-socketio) have written it
 
@Downgoat this is actually kind of nice
 
@mınxomaτ I don't know nginx or apache well enough to know what you mean.
 
@mınxomaτ everything just works as long as you stay within the framework. I don't make any guarantees as soon as you start doing other stuff
 
Anonymous
That's why I dislike .NET: it's a monolithic framework. It's very difficult to replace parts of it.
 
1:55 AM
@Downgoat That works as well
 
@ConorO'Brien Better if the S wasn't flattened (cropped-looking) at the top/bottom, but yeah.
 
@ConorO'Brien nicer than?
5 mins ago, by Downgoat
user image
;_;
 
well idk
 
@Mego if the framework is simply the standard of communication, then people have to write modules for it. It doesn't just work with any arbitrary module.
unless, of course, you have adapters, but that requires extra code, aka, not cohesive
 
Of course, but modular means you can replace individual components
 
Anonymous
1:56 AM
@NathanMerrill Sure, but that's a lot better than a monolithic framework that is everything-or-nothing.
 
@Mego right, but even django allows you to do the same. A great example is all of the templating languages they allow you to use
 
Anonymous
@NathanMerrill Adapters aren't necessarily incoherent. If you view an adapter as a module, it does exactly one thing: provide an interface for a module to connect to the framework.
 
@ConorO'Brien ;_; wait VSL is Downgoat's now?
 
I consider adapters to be extra cohesive :)
 
@ASCII-only you weren't here; had you have been, I would have said "their"
 
1:58 AM
they glue two non-cohesive parts together
 
@ConorO'Brien cheddar is my new scripting lang, not VSL :P
 
@Downgoat kokok
 
VSL = Very Sharp Language = Cheddar
Confirmed
 
I'm going to be writin a scripting language soon :P
 
1:59 AM
once I learn enough C++
 
@ConorO'Brien There are like 6 collaborators
 
@ConorO'Brien Why is C++ a requirement
 
@ASCII-only I want to write it in C++
 
Anonymous
@NathanMerrill I'm starting to think you aren't using the same meaning of "cohesion" as I am.
 
2:00 AM
Non-esolang, right?
 
Why dont you just contribute to VSL if thats the case? :P
 
@ASCII-only yeah. It's going to be basically J, but usable.
 
@Mego is there a programming-specific definition for cohesion?
I just think "how well the parts fit together"
 
Anonymous
In computer programming, cohesion refers to the degree to which the elements of a module belong together. Thus, cohesion measures the strength of relationship between pieces of functionality within a given module. For example, in highly cohesive systems functionality is strongly related. Cohesion is an ordinal type of measurement and is usually described as “high cohesion” or “low cohesion”. Modules with high cohesion tend to be preferable, because high cohesion is associated with several desirable traits of software including robustness, reliability, reusability, and understandability. In contrast...
 
Then use bison, if it isn't an esolang either use C++ and Bison, or PyPy
 
Anonymous
2:02 AM
Basically "how much do the parts of this module relate to each other?"
 
Anonymous
It's not "how well the parts fit together" - that's more like coupling.
 
ok, let me take this another direction: A linux distro does many things. You could consider it un-cohesive. However, I'd consider it cohesive because different things are in different places
I believe the same to be true for Django. You're right that it does a lot of things
but it still has a good separation of tasks
 
Anonymous
I don't consider the cohesion of a linux distro because it is made up of many cohesive modules which interact with each other via the kernel and other modules specifically made for intermodule communication
 
Anonymous
You can switch out the individual modules in a Linux system (like zsh instead of bash) with few problems because it is modular
 
for example, routes.py and views.py. Flask thinks its a good idea to mix url routing and view calling, where Django keeps them seperate
 
Anonymous
2:06 AM
Cohesion is a module-level quality. Coupling is a system-level quality.
 
I'd definitely consider that uncohesive
 
Anonymous
@NathanMerrill I don't understand where you got that incorrect impression
 
the fact that you put annotations on the views themselves
20 mins ago, by Mego
@NathanMerrill But @app.route! Routing is done via decorators!
 
Anonymous
@NathanMerrill You can, but you shouldn't. You define the routes via decorators on functions, which then call code that creates the view, and returns it.
 
oh, I'm mixing up terms
routing is in the same spot as the controller
I do think Django made a bad decision calling their controller file "views.py"
but regardless, django does separate the controller from the routes
 
Anonymous
2:09 AM
The routing functions should manage the request parameters, and then hand off processing to the controller code. They then take that result, call the view code, and return the resultant view.
 
Anonymous
You absolutely can (and should) separate the controller from the routes.
 
Anonymous
But you also can make it entirely monolithic.
 
so, you would recommend a flask project to put all of the @app.route annotations in a file that then calls the actual controller functions?
 
Anonymous
That's probably the best way to do it
 
Anonymous
Though maybe s/file/module
 
Anonymous
2:12 AM
(since modules can and often do span multiple files, thanks to Python's subpackage and submodule systems)
 
sure :)
 
Anonymous
Flask has a lot of flexibility that Django lacks
 
@NathanMerrill the app.route should also handle all the url request and templating/url response stuff before/after passing it on
 
@Maltysen what do you mean templating?
like defining which variables gets passed to the template?
that's the controller's job, not the route's
@Mego I'm still not convinced this is true. I haven't tried breaking out of the Django framework beyond using a different templating system, but the controller modules are separated from the model modules, which are separated from the routes modules
 
@NathanMerrill srry didn't mean templating, meant url response object
 
2:15 AM
what's a url response object?
 
Anonymous
@NathanMerrill HTTP 418
 
@NathanMerrill you can just return a string, but you can return more complex responses
like redirects, headers, attachments, json, etc.
 
while I would agree that redirects are the job of the router, I think that any other http status would be the job of the controller
though I'm not sold on that
 
well I think that really depends on the specifics of the situation
 
and attachments/json is definitely supposed to be added by the controller, not the router
 
2:17 AM
@NathanMerrill not the json part
 
its content!
 
the controller can return a python object, and with Flask's magic decorators, the router can jsonify it & apply correct headers
 
oh, the act of jsonifying an object.
still, on a high level, there's no difference between jsonifying an object and HTML-ing an object
they are both a "view"
and I don't believe that routes define which view you use :)
 
btw, does anyone know if socketio maintains packet order?
 
2:34 AM
@NathanMerrill I think I've made some really good progress on a mathematical model.
 
awesome!
I forgot to reply to your last message :)
 
It was really difficult. (I did everything empirically.)
 
so, what's the input, and what's the output?
I pass in a list of rankings
and a "final ranking"?
 
Hypothetical situation to demonstrate the progress I've made: there's 5 bots in a competition. The average score of each bot is {0,1,2,3,4} respectively, with a standard deviation of 1 for each. Let's say that each tournament consists of X samples, which are averaged to determine the ranking. If I were to perform two tournaments, what's the expected distance between results, as a function of X?
Understand this situation before I post the answer?
 
I don't understand "expected distance"
 
2:39 AM
The "distance" between two sets of rankings, is the sum of the squares of each individual's difference in ranking.
 
in this case, expected difference is very closely given by "y = 5 * (b^x + b^(4x) + b^(9x) + b^(16x))" where b ~= 0.74
That was very painful to determine.
 
I'm not sure where that equation came from, but I'm not really concerned about that :)
 
The problem is determining what numbers to use inside of the parenthesis.
Because, for any given real-world situation, those values aren't going to be the same.
1.0 : 5.670874 : 5.572458412915903
2.0 : 2.97969 : 3.2100623789384635
3.0 : 1.88415 : 2.162414580404068
4.0 : 1.288296 : 1.5398543997820493
5.0 : 0.91701 : 1.121632675138673
6.0 : 0.666924 : 0.8246681097193921
7.0 : 0.489006 : 0.6086541219901653
8.0 : 0.362876 : 0.4499242495294422
9.0 : 0.268834 : 0.3328000733750204
10.0 : 0.201036 : 0.2462289125297099
15.0 : 0.049408 : 0.05463190780571652
20.0 : 0.012644 : 0.012122840889412057
30.0 : 8.56E-4 : 5.969275127744501E-4
50.0 : 6.0E-6 : 1.4472914313448682E-6
sample size : actual value : prediction based on my empirical fitting above
so not perfect, but it's something along those lines
 
2:50 AM
I'm not sure what that table means
labels?
 
1 min ago, by PhiNotPi
sample size : actual value : prediction based on my empirical fitting above
 
oh duh
lol, was looking above
wait, so you did a curve-fit for the above equation?
or did you derive it?
it looks derived
 
@Downgoat . is totally operator, snek lets you overload
 
no... I curve-fit for a smaller competition (3 competitors) and mathematically extended it for larger numbers, and it works.
which is important because it means that the mathematical pattern I saw is something real
 
@Maltysen snek is stupid :P
but seriously wat
did python people not know parsing 101
WHY WOULD OYU MAEK . AN OPERATOR I DO NOT UNDERSTAND
 
2:56 AM
hmmm...I'm still not sure how I apply this. I mean, I can easily take each of the rankings and find the average sum-of-squared-difference
but once I find that, how do I get to the "actual" value you have listed?
 
The "missing step" is to figure out how the exponents/bases are distributed in "real competitions."
I know that the curve follows a pattern of y = a * (b^x + c^x + d^x + e^x) etc. but the question is: given a real competition, how to easily determine values of b,c,d,etc. that work well enough.
 
Anonymous
@Downgoat Because flexibility
 
Anonymous
Let's say I wanted to implement lambda calculus in Python
 
Anonymous
I would want to overload . for lambda calculus functions to mean abstraction instead of member access
 
Anonymous
Or making a tacit language - I would want . to mean composition
 
3:07 AM
@PhiNotPi one other question: This is a confidence metric, right? But what are we making confidence of?
do I pass in a ranking, and it passes a confidence that that ranking is correct?
or is it simply a statistic of how variable the rankings are?
 
Here's the goal: pass in a bunch of rankings using small sample sizes. It determines the relationship between sample size and variance, somehow (curve-fitting with some template equation). Then, you pass in a single ranking of a much higher sample size, and it tells you how likely it is that the ranking is correct.
 
so, theoretically, I could pass in each of my games as "data", then pass in my MAM-calculated result, and it could tell me how accurate it is
 
To make it even smarter, it should be able to do the curve-fitting when you continuously pass it updated rankings as the tournament is being held.
To be more specific, you would have to pass it MAM-calculated results of a small sample size, so then it could tell you how accurate a large-sample-size MAM result is.
If this idea actually works (which it might not), then it would have been extremely helpful with my NanoCoreWar KOTH.
 
why do they have to be Mam-calculated results?
(for the small sample size)?
 
Since in that one, each pair-up was played several thousand times. I could run each pairup once, compute overall rankings, run each pairup a second time, compute updated rankings, etc. several times. This program should (eventually) be able to determine the rate of convergence, and tell me when I can stop performing matches.
@NathanMerrill It will get fed rankings as data, not game results as data.
 
3:22 AM
there's no difference though?
 
What do you mean?
 
I literally use the same object for game rankings as tournament rankings
both of them spit out an ordering of players
 
The ranking algorithm (MAM in this example) is what combines games (sample size 1) to an aggregated ranking of a larger sample size.
And I need a variety of sample sizes to determine how sample size is related to accuracy.
 
oh, I understand. You're measuring convergence, not accuracy
 
In some cases, increased sample size does not increase accuracy at all.
And... I believe accuracy can be somewhat estimated from the rate of convergence.
 
3:37 AM
Let's say that I have some data, and I want to fit it to y = a * b^x + c * d^x. What's the best/easiest way to do that?
As an added bonus, I know the sum of a + c.
I need at least 3 data points, preferably more.
 
This may be really random, also it may already have been said, but another problem we have with getting top golfers to join us is that we generally only have one competitive submission per language
 
@Downgoat it's really cool. it calls __getattr__ which you can override, but it also calls __get__ on the thing being got, which is how the magic self works
 
@Downgoat in most languages [] is subscript operator
 
@Downgoat also properties which are super-duper useful
 
4:24 AM
Is @Adnan here
Guys
 
Doesn't look like it
 
In 05AB1E, how would i turn ["1234", "123", "12", "1"] into [["1", "2", "3", "4"], ["1", "2", "3"], ["1", "2"], ["1"]]
How would I loop through a list in 05ab1e
ope
finally got it
its
€S
 
4:52 AM
Why us sell army and weapon to Arabic Saudi ? Or Am I not well informed? The distance in mathematics has a definition and propriety
 
5:22 AM
You have just expressed three disjoint thoughts. Can you clarify?
 
0
A: Sandbox for Proposed Challenges

DJMcMayhemCan I sweep the mines? code-golf grid game Minesweeper is a popular puzzle game where you must discover which tiles are "mines" without clicking on those tiles. Instead, you click on nearby tiles to reveal the number of adjacent mines. One downside about the game is that it is possible to end...

 
@NewSandboxedPosts Thankyou!
 
 
1 hour later…
6:45 AM
@Everyone help what should a golfing language do for for loops with a negative variable e.g. for(-3)
 
Iterate in the reverse order it would for for(3)
 
7:12 AM
I put in the link to repo I stupidly forgot, and added new snippet
2
A: Showcase your language one vote at a time

Destructible WatermelonDotsplit (whoops I forgot the link before :p) (pair of) Factoid(s): Dotsplit is a mostly syntax free language, with odd builtins (kind of like mathematica), but is currently in it's infancy. Dotsplit is named after the tokeniser used in the interpreter: str.split(). This is a weak tokeniser, b...

 
7:43 AM
 
8:07 AM
> tfw you name your blankets after top stackoverflow users
 
@ASCII-only if your language is capable of inverting (some of) its own functions, then iterate the inverse of the function body 3 times
 
help me design a golfing version of ruby
 
November 15? Plenty of time :D
 
Installing Allegro for C++
 
> Can you design a language whose programs are games of chess
^ I think @Sherlock9 would be interested in this one (link for Sherlock)
 
8:17 AM
@DmitryKudriavtsev ruby.janlelis.de/…
 
@ASCII-only that one is basically taken verbatim from my list of language ideas :D
 
@ASCII-only Indeed. No idea where to start though :D
 
That would be one heckuva programming language
@Sherlock9 Can I help?
 
I wonder if submitting an existing language is allowed (also ;_; looks like it's Argentina only)
 
8:18 AM
If so, what lang is it going to be written in?
 
@Sherlock9 Yeah
 
@ASCII-only it's not
> The contest is open to the general public: professional programmers, amateurs, students, and anyone else interested, without regard of their nationality or place of residence.
 
Wait wait wait what's the contest?
 
Oh TIL
 
8:19 AM
@Qwerp-Derp bold link a few messages up
 
Ah
Should we do Chess? It seems like an interesting language
 
@Qwerp-Derp I'm also in the middle of midterms and one or two other projects. Designing a novel language when I have never designed any language before (not fully, anyway) would be a bit of a handful
 
I can use the ways that @ASCII-only has shown me when writing Logicode
to make Chess
 
IMO Retina/Jelly/V are worth submitting, maybe Charcoal just to name a few
@Qwerp-Derp You should probably use Charcoal's parser, it's tidier plus once you know what does what you can adapt it to Logicode as well
 
Ah
Nah Chess is fun
 
8:24 AM
Anyone know a good compression library for non-word strings?
 
@Qwerp-Derp It is a month away though. If you come up with something or vice versa, let's help each other, yeah?
 
Sure, but I've got exams in about a month or so, and I might need to start revising
 
Ah, fair enough. Good luck to us both
 
@ASCII-only Where's the parser?
@Sherlock9 Ditto!
 
@Qwerp-Derp Repo, main parser is near the bottom of charcoal.py, any file that has only one dictionary is probably a parser or a dictionary of lambdas used to process parser output
@Everyone how useful is RLE/shoco for strings for an ASCII-art-focused language
 
Anonymous
8:30 AM
@ASCII-only Shoco will be useless
 
Depends on how many challenges use English words
 
Anonymous
@MartinEnder I'll write a language called Tearless
 
does game of life count? :)
 
Anonymous
No, GoL has been known to be TC for a long time :P And unless you're John Conway, you didn't design it
 
Anonymous
8:36 AM
Though Phi might be interested in submitting QFTASM and/or Cogol
 
8:49 AM
@Mego is that a pun I don't get?
 
Anonymous
"Don't Cry for Me Argentina" is a song recorded by Julie Covington for the 1976 concept album, Evita, and was later included in the 1978 musical of the same name. The song was written and composed by Andrew Lloyd Webber and Tim Rice while they were researching the life of Argentinian leader Eva Perón. It appeared at the opening and near the end of the show, initially as the spirit of the dead Eva exhorting the people of Argentina not to mourn her, and finally during Eva's speech from the balcony of Casa Rosada. Covington was signed by the songwriters for the track, based on her previous work in...
 

« first day (2089 days earlier)      last day (2755 days later) »