« first day (2412 days earlier)      last day (2428 days later) » 

10:00 PM
@MagicOctopusUrn Under Windows (hopefully soon under Linux and macOS too) you can just type ]chart mydata to graph anything in an interactive GUI. You can then export as image file or SVG, etc. Generating the files works on all platforms already.
@MagicOctopusUrn Well, APL itself is almost our only "product", and we license it for use by very large enterprises (and very small ones too).
 
@Adám just wondering, what your biggest consumer base is? Is it AI? Accounting? Insurance? Etc... Cool if that's not public domain, but that's been my major question; who is using APL to the fullest extent? In what domain.
 
@MagicOctopusUrn I'm pretty sure finance is Dyalog's largest customer section. E.g. SimCorp and AMECO run on Dyalog APL. However, production management is also a major APL sector, with representatives like Volvo Penta, and also one of the world's largest oil companies (which does not allow us to mention their name for fear that their competition will pick up APL…)
@MagicOctopusUrn I've got to go to bed soon, but see case studies and reference customers for more of that type of info.
 
6:11 EST: My freecell solver is stuck on the part that takes it 300k turns
 
@Adám - oh interesting that there's a font-to-back integrated system in APL... kinda what I've done (in-house) over the years using Python. Am about to start at an exchange working on their batch jobs also using Python. Python is everywhere in finance that is not real-time / low-latency where C++/Java tend to take to the stage; but I'm yet to encounter any APL.
 
10:16 PM
@JonathanAllan Most companies keep their APL to themselves. They train their own workers, and guard their code fiercely. Btw, have you heard of Py'n'APL, the new APL-Python-APL bridge written by PPCG's very own marinus?
 
Yeah most finance companies keep most stuff to themselves :p I had not heard of it, no...
looks interesting :)
 
@JonathanAllan Sometimes Dyalog APL is actually faster than optimal C – which is of course impossible as Dyalog APL is implemented in C… However, those financial institutions that need real speed use K, which is also a type of APL, but with everything removed except that which is essential for financial databases. K is usually faster than C, despite also being written in C.
@JonathanAllan One of the world's largest providers of financial services have their system written in Dyalog APL, but again, I'm prohibited from saying which company that is.
 
HFT shops often use C for their algorithms from what I've seen - and they are the ones who are really battling regarding speed of action (so much so that they run specialised hardware too)
"financial services" covers a lot. JPM were using Java for their high frequency FX trading (I sat right next to them).
They are pretty big :p
 
@El'endiaStarman Have you considered open-sourcing TNBDE? :3
 
But then they'll have other tech in other places
 
10:24 PM
@JonathanAllan Right, but writing in C has a huge downside: If the market changes, and the models need updating, it takes much too long to rewrite and recompile compared to rewriting in K. The K interpreter fits in cache…
 
they do gpu stuff & fpga...
yeah well FPGA is the extreme of that in terms of making changes
Much of what FX needed could be achieved with Python surprisingly (yes I know it is slow!)
 
@JonathanAllan Morgan Stanley wanted APL to be customized to their needs, so they created a dialect called A+. Later that became the basis for K, which was then spun off into a separate company, Kx Systems.
 
Oh right, how long ago was that do you know?
 
My government teacher's idea of a good lesson was to turn on CBS news (90% sure it stands for Certainly Biased Sources) and have us write 3 sentences for each issue they talked about.
In the end after 50 min of news 6 sentences were written
 
Ah I see Kx was founded in 1993
 
10:31 PM
@DJMcMayhem 68 bytes
 
quite a while ago
 
I always find it amusing to contrast the flashiness of Kx's homepage with their absolute minimalist K language homepage.
 
I do know that the project I worked on over at GS, JPM, BAML was a flop at MS though...
Oh those big-scroll web pages - I'm not too keen on them
I will probably be getting onto doing some hadoop...
 
so...I got chrome 61...it's really not autoscrolling in chat
 
:(
rollback
 
10:36 PM
impossible :p
you don't choose if you want to get it it's automatic
 
ah I think you /can/ if you google it
 
it's disabled :(
 
Does that flag enable some experimental ES features as well I'm guessing?
can reproduce on Chrome 61 (dev). enable-experimental-web-platform-features is offbadp Jul 2 at 13:09
 
and...don't the colors look a little different on chrome 61?
 
10:41 PM
yeah good luck on ubuntu
 
sounds a little involved though!
 
yeah you can't even click the reply arrow thanks to chrome 61
 
There is somewhere in Chrome to switch between betas and alphas but not quite sure where it's hidden
 
chrome flags?
the issue is on SE's side not on client side
 
10:42 PM
@EriktheOutgolfer no you can (i can?)
 
your reply goes too far away to be able to click it
 
Is chrome borked completely?
 
more like outdated stack exchange javascript
 
oh ok
 
dev.opera.com/articles/fixing-the-scrolltop-bug this guy called it in...2015? wtf?
 
10:46 PM
@EriktheOutgolfer Scroll down to "Switch to a Different Channel of Chrome in Ubuntu" and you'll be able to get back to 60 I believe... howtogeek.com/howto/21282/…
 
one should really report the bugs to the SE team
 
seems outdated...
@Mr.Xcoder it's already posted on meta as of 10h ago
 
could be
 
@betseg really, wow
 
@JonathanAllan I have the google-chrome-stable channel
so that doesn't really help
 
10:49 PM
:(
 
I guess it's solely SE's fault then
 
So my dad calculated that Social Security tax (if you get paid out in full) will end up costing you about $27k a month. If you invested the amount you pay in social security (assuming you max it out) according to projected F&B500 growth rates you would be getting ~30k a month instead of the social security 3k a month
Our government cannot manage money at all
 
0
A: Max power of 2 such that 2^m divides n. (The ruler sequence)

VaelusMath Student, 47 bytes evaluate {max{i∈ℤ>-1:2ⁱ∣n}:n∈ℕ≤N} Note: some math students may hang indefinitely upon receiving the program. There is much debate about the cause of this bug.

Um
 
HAHA XD
 
@H.PWiz Nice! Here's what I had: Try it online!
 
10:52 PM
I think it's "Mathematica Student Version"
 
Click the link
THey mean math student
 
I feel like I shouldn't but I upvoted since it was too good
 
@DJMcMayhem I also posted a 66 byte one in the third stack.
 
It may work in mathmatica but idk
 
10:53 PM
Is that considered valid?
 
@2EZ4RTZ probably not
 
@Mr.Xcoder Idk but it is really funny
I vote to keep it
 
I vote not a chance
 
I vote Mr. Xcoder
 
Pls downvote and delete
 
10:54 PM
Or we can wait and find out if it was a joke and the code works
 
good luck with that
 
@2EZ4RTZ There is no such language I believe
 
Maybe we should explain that in order to be a valid submission it must run on some existing interpreter (and "a math student" does not count)
 
I ran on me
 
10:56 PM
Wow, I just experienced the worst fail in my entire life....
 
brb typing
 
@DJMcMayhem I am going to leave TNB for 10 min. Plz kick if I join back
 
> Apparently other people didn't find it as funny.
 
I stayed up until 1 o'clock waiting for a tennis match to start. Then, I stayed up until now (2 o'clock) assuming they're just being slow. I look at the schedule... The match starts in 3 hours... facepalm
 
10:59 PM
woah i didnt notice you were romanian
 
I have literally wasted 4 hours of my life
 
rip
 
@betseg Wha't that woah about it :-p
 
dunno, thought you were american or something
 
No, why would you think that way?
nvm
 
11:00 PM
because a lot people here are from america? :P
 
Dangit, I didn't get a chance to explain why the answer wasn't funny
 
@ETHproductions I think the poster understood
 
@Mr.Xcoder btw you being close to me is woah-y
 
We beat you to it, ETH
 
@Mr.Xcoder as american I'm pretty sure USA is only country in the world :P
 
11:01 PM
@betseg Oh, you live in the surroundabouts.... Now I wonder where
 
I think turkey
 
Probably
 
my profile: "some weird place that has the same name with a strange bird"
 
even though turkey isn't a strange bird
 
yeah, definitely turkey
@EriktheOutgolfer <insert look of disapproval>
 
11:03 PM
@EriktheOutgolfer cant even fly! what kind of bird is that
 
we make well known turkey ham out of it...?
 
Also, Arda Ünlü is pretty self-explanatory
 
@betseg then what about chicken or peacock?
 
@EriktheOutgolfer theyre also strange
 
^
Sad to say, Mego is "strange" now :DDD
Apparently they're a bird and can't fly
 
11:04 PM
they can fly...but under the sea
 
> flying under the sea
 
yeah it's also known as swimming btw
 
I hate random downvoters from the very bottom of my heart...
And they undownvoted...
 
Possibly accidental, especially if it only lasted a few seconds
 
yes...
CMC: Given two lists, determine if one is a prefix of the other one
 
11:08 PM
can we assume the shorter one is first?
 
@ETHproductions Yes.
You can assume that.
 
ES6, 30 bytes: (x,y)=>x.every((g,i)=>g==y[i])
 
@ETHproductions Does japt have prefixes? can't find 'em in the docs
+1 for the optimal python solution
 
If it's a string you could do å+
 
11:11 PM
There's no super-easy way for lists though
 
Oh, that should be a built-in
 
Yeah
Japt, 10 bytes: d@¯YÄ ¥V+P
 
@Mr.Xcoder Bash, 27 bytes: [[ $2 = $1* ]]&&true||false outputs via exit code
 
(there's also no super-easy array equality check in Japt :P)
 
11:13 PM
really >.>
I want to give a +1 so... anyone up for finding the optimal solution in Python?
 
@Mr.Xcoder Husk 3 bytes
 
Prefixes, count the occurrences?
 
Japt, 7 5 bytes: eȥVv
 
TY DJ
 
Wow the 5-byte version is strangely aesthetically pleasing
 
11:18 PM
eEYVv
 
@ETHproductions I have an algo... maybe it would save bytes: the first array == the second one up to the length of the first one?
 
oh nice idea
 
@Mr.Xcoder ... Yes what a great original idea ...
 
@H.PWiz What are you saying?
 
@Mr.Xcoder That would be ¥îV, except ¥ (==) doesn't work on arrays
 
11:19 PM
>.<
 
So ¥îV +P for 6 bytes
 
Have you seen my answer Πz=
 
RProgN2, 10 bytes. ;-; iy=x³y‘er*
 
@H.PWiz I didn't bother to look in the docs as to what they are.
So no
 
@Mr.Xcoder fair enough.
 
11:21 PM
+1 if you outgolf me 24 bytes (in Python)
 
9 bytes, infact. iy=³y‘er*
 
Always nice when you find out you can just remove a char
 
@Mr.Xcoder C, 48 (thanks for the idea, @Mr.Xcoder!)
 
nice
now do it in C++
oh, look the tennis match actually started
 
@Mr.Xcoder ok
@Mr.Xcoder which one
 
11:25 PM
@betseg The US Open one, Keys vs Kanepi
facepalm tries to read a whole line with std::cin
 
@WheatWizard I will find a class like that and take it and do everything in brainfuck/flak/miniflak
 
gtg bye
 
11:39 PM
me too, cya guys later
 
:(
 
@ATaco userscript request. Poll board. When a strawpoll link is posted it goes on a board of the 5 most recent strawpolls and displays on the side or smth
 
Hmm, I'll look into it.
 
11:59 PM
@Downgoat Looks pretty to me.
 

« first day (2412 days earlier)      last day (2428 days later) »