« first day (771 days earlier)      last day (3674 days later) » 

12:17 AM
Can someone please tell me the reasoning behind "therefore" in the second line of the proof here: en.wikipedia.org/wiki/…
 
12:30 AM
@WuschelbeutelKartoffelhuhn pidgeonhole principle?
 
 
1 hour later…
vzn
1:43 AM
6
Q: What should be included in a freshman 'Mathematics for computer programmers' course?

Brian RushtonMany universities are changing up the way that they teach math service courses. 1-3 semesters of calculus and maybe a course in linear algebra are often included in majors (such as computer science) that will not use the majority of the material. However, these courses are often justified by say...

 
 
3 hours later…
vzn
5:10 AM
0
Q: list of cool research project topics

vznover 2yrs have noticed that ppl asking about what to work on for research projects (undergrad, grad, phd) is a common query around stackexchange. these questions are almost invariably quickly closed on main sites. this seems a bit of a shame to me. for most of us, we are all in CS for personal re...

....
mass meta split!
246
Q: Operation 'Split, All The Metas!' Shall Commence On April 16, 2014

Tim Post Update 2014-04-17 04:47 UTC: Stable, with some odds and ends to clean up The main MSO/MSE sites are stable, but there are plenty of corners left to sweep. For details, see: The Status Of The Split Not all of them mind you, just Meta Stack Overflow into Meta Stack Exchange. I'll have yo...

 
 
2 hours later…
7:16 AM
Ok, you insist on using silly abbreviations, not capitalizing words and generally breaking the conventions of the English language. Fine. But do it in your blog or in your volumes of poetry. We are here to communicate. That means using our common language, English. Making mistakes is not a problem, but when you deliberately write in order to be hard to read, I'm not going to spend time trying to decipher what you wrote. If anyone cares to edit this question and translate it into proper English, and it turns out to be on-topic here, I'm all for reopening it. — Gilles ♦ 24 secs ago
@vzn I've told you before: if you want to communicate with others, don't speak in a lingo of your own.
 
 
4 hours later…
11:22 AM
Is there some way of running queries over stackexchange data? I'm interested in quantifying overlap between tags on various sites (cs and scicomp, dsp and electronics and cv, ...).
This is for a meta.se question I am composing requesting the ability to cross-post a question to several sites simultaneously.
 
12:22 PM
@WanderingLogic yes: through the API, or through the data explorer. Which one to use depends on what you want to do. Neither method makes it easy to do queries overlapping all sites, but queries overlapping O(1) sites can be done
@WanderingLogic this has been proposed many times and shot down, make sure to read the existing work before proposing something
 
12:56 PM
@Gilles thanks!
@Gilles I have been reading some related meta posts. That's why I need to collect data.
SE staff responses are invariably along the lines of "there is not possibly any overlap between sites."
 
 
2 hours later…
vzn
3:25 PM
gilles re kneejerk closing the meta question, there is nothing unclear about it
2 hours ago, by Wandering Logic
SE staff responses are invariably along the lines of "there is not possibly any overlap between sites."
lol!
ridiculous
wanderinglogic, crossposting is an interesting idea but would bet $x that it wouldnt be implemented in a million years
the se architecture is pretty mature & afaict major chgs havent been made in years
eg migration is an older mechanism that hasnt chged much.
Realz, interesting! thx for link
did you see my recent blog on games? cited you
 
 
1 hour later…
4:51 PM
"Each site is focused on a specific topic area. If you have a question you feel is too ambiguous, either re-think the question carefully, or [get it migrated].

99.99% of questions people have labeled as being applicable cross-site have been proven to be valid on a single site if written properly and thought through. SE is not a wild west for questions; a question needs to be worked on to be worthy, and if worthy, it will target a specific audience."
=D
The problem I'm really concerned about specifically is the rapid proliferation of applied computer science betas either already going or being proposed.
For a lot of these there will be substantial overlap between the proposed new site and some combination of cs and so (for the programming questions)
It's a reasonable argument that it is difficult to build a cohesive community when your question base needs to be divided between two existing sites.
So you get proposals for operating systems
gotta go, but the machine intelligence community keeps trying
the scicomp community and dsp community and ee communities all have overlap with us
and there were several mostly overlapping "computer engineering" type proposals a while back.
*********************
so the real question is: how can we allow subcommunities to form effectively?
Area51 + no crossposting clearly ain't it
 
vzn
5:16 PM
hi wanderinglogic.
another issue is that some sites are long in beta and are not ever reaching "critical mass" eg this one. so why are there new cs-related proposals in area51?
know what that suggests to me... that the existing sites are too inflexible/conservative/restrictive in their charters.
yes stuff like AI doesnt fit too well in existing sites (despite a tag for it), was disappointed not to see it succeed in area51.
am starting to wonder if the best idea is some kind of capability of reorg based on tags, where a site could accumulate tags but then say "lets push these tags somewhere else"... there is some discussion of that with the new stackexchange meta site. [gilles himself proposed that in a meta response]
imagine if all the ai tags across existing sites suddenly got re-org'd into a new site... hmmmm
yes dont really like the area51 mechanisms myself much, its very difficult to build communities there... :(
maybe purposely so.
area51 is yet another se ghetto-like area.
 
@vzn: I dislike area51 for perhaps the opposite reason: it makes it too easy to try to start a community instead of encouraging peopel to join existing communities.
 
vzn
maybe ppl have attempted to join existing communities & get rejected
se culture is restrictive & compartmentalized
at times
 
Perhaps. For the OS/Compiler/PL proposals you can see in the "Followers Active In" stats on the side that CS is not in the top 6.
 
vzn
little or low toleration for diversity
 
I think it's more that they didn't know that CS exists.
(And certainly don't know that scicomp exists)
 
vzn
5:23 PM
yeah well thats the issue of site promotion which weve discussed lots on chat
 
But we (here on CS) can be kind of hostile
 
vzn
exactly, but that is true of every se site it seems
dont see a solution but think you are asking the right questions.
se suffers from "when you have a hammer everything looks like a nail"
there are lots of hammers in se...
downvotes, closing questions, banning users, etc
deleting comments/questions....
poof
migrations...
editing users questions to strip out stuff....
 
I think we do all that stuff, but mostly to neophytes.
 
vzn
if you look at the mass users on a se site, the rep is low. therefore essentially most users are neophytes.
my pt is that how neophytes are treated determines whether they come back.
 
I meant neophytes == undergraduates
 
vzn
5:28 PM
if their 1st experience is bzzzt, zap then they are unlikely to come back.
 
I agree that if you zap someone they are unlikely to come back
And if bringing up the newest questions link gives you a page with questions mostly on hold you are going to be somewhat put off.
Not that I actually think we are particularly overly aggressive about closing questions.
Most of what gets closed is really bad questions.
But what I'm particularly worried about is how we attract a critical mass of experts in applied cs.
We are really strong on algorithm design and analysis here
and also strong on PL
 
vzn
agree most closed questions are borderline but doubt it has much to do with attracting experts.
 
but there aren't many of us here who publish in conferences like CGO, ISCA, MICRO, OSDI, SOSP.
 
vzn
mods argue that borderline questions make experts flee, but kind of doubt that.
its one of those articles of se faith circulating among the masses....
the so called shiny quality questions
anyway if a cs site is stuck in beta for a long time, why should a CS subtopic proposal be raised in area51... agree it makes no sense
 
I don't think the problem is too many bad questions
but there is a lack of quality questions in applied cs
 
vzn
5:39 PM
see it more in terms of attracting quality users rather than questions
 
yes.
but how do you do that?
 
vzn
the $1M question.
heres a crazy idea.
every time two people on the site disagree about something, it loses )( that much momentum.
find a way to agree....
se seems at times like a disagreement machine....
my zen question, is it the software design or human nature?
or both?
just some crazy musings from the court jester.
 
vzn
funky
that reminds me of a cogsci question
high voted
21
Q: What is the psychology behind trolling?

zergylordFor those new to the internet, trolling is an activity where one person intentionally tries to upset other members of the same community, presumably for entertainment. This has been informally addressed in the media, normally positing that the anonymity of the internet gives rise to more extrem...

its very strange though, cs enrollments are up at colleges at times.
the so called "social network" effect (re movie, seemed to have a bump on CS enthusiasm)
its too bad stats are not kept on failed area51 proposals, its rather staggering.
 
That's a good question.
MAKES ME WANT TO PICK A FIGHT!!!!
 
vzn
5:47 PM
lol
 
vzn
subversive & likes fighters =D
yeah area51 is something like a big semisticky flytrap
 
There have been some resounding successes (outside of cs)
 
vzn
on area51? like what?
anyway se has never been a lot about building new communities its more about trying to get ppl to fit into existing boxes....
 
I think Gaming, English and Cooking.se all came off Area 51.
 
vzn
5:50 PM
actually almost all sites came out of area51 afaik.
100+ sites now.
it is impressive.
se mgt seems to have no response for many sites stuck in beta though.
 
I think we're closer than we think.
 
vzn
its just endless "wait and see"
 
I don't know how to search in this chat room
 
vzn
upperright search box.
the search is indeed tricky.
 
but there was just a post a week or so ago where they said that the only remaiing bottleneck is getting a graphic designer freed up for us
 
vzn
5:51 PM
also click info tab.
yeah shog9 might have been saying that.
am doubtful the site is not graduating because of the graphic designer.
gilles was talking a lot about the design iirc.
like your team spirit yay go team
thx for the info re parallelism, have found it highly fascinating subj for yrs.
sounds like youre working on some cutting edge stuff.
 
vzn
yrs ago knew guy who went to work for intel, had phd iirc
an amazing company
do you do any research there?
 
Not much research. More prototyping and advanced design.
I never much liked writing papers when I was an asst. prof.
(which is probably why I'm not an asst. prof. any more)
 
vzn
lol
what languages are you using?
 
I get a really good intern now and again who will write a paper with me.
 
vzn
5:58 PM
cool :) looking for good interns myself in meta lol
 
Languages: we do most of the tool dvpt in C++
Languages we analyze gets weird
Pin handles machine code
And we can handle the standard debugging symbols on Windows and Linux
 
vzn
←have debugged tricky race conditions over yrs & answered few questions related to it/parallelism etc
 
But on the GPU side it's a mess.
 
vzn
re area51 there is an interesting feature where when the proposals fail, existing mods are allowed to "poach" the questions.
 
We're trying to work with OpenCL, Intel(R) Cilk(TM) Plus, OpenMP 4 simd pragmas
 
vzn
6:02 PM
gilles did that a bit with the AI proposal.
 
I saw one of those the other day.
 
vzn
not sure how many questions came out of that.
guess its easy for them to "skim" the highvoted questions.
so all is not lost.
 
An AI question that got migrated from the failed AI site
 
vzn
yeah.
se basically does not try to support "small" communities much.
its all about scale.
 
I wish we could get the people to migrate in with the questions/answers
 
vzn
6:03 PM
have seen shog9 even complain about tcs.se being too small in a meta post.
would like to see se experiment with smaller, more dedicated/passionate communities more.
have collected tons of refs on se culture, planning to blog on it sooner or later. blogged on it once, re the hair-trigger closing tendency cross-sites.
like facebook, "its complicated"
 
Math Overflow gets 42 questions a day
 
vzn
yeah. think thats kind of cool. think the slow moving sites (cs.se, tcs.se) are kinda less interesting.
 
TCS gets 8.5 (which is much better than it was a year ago.)
 
vzn
math.se is another huge success story, moves fast, relatively new site.
 
I think Math Overflow and Math are the golden standards
 
vzn
6:06 PM
MO is quite the phenomenon and largely inspired tcs.se.... tcs.se is still a MO "wannabe".
never will reach MO scale.
 
There's a new Mathematics Educators beta that seems to be doing quite well.
 
vzn
the MO experts are incredible. real top leaders in mathematics.
 
Maybe Mathematicians are nicer people than Computer Scientists
 
vzn
yeah have dinked around a bit on the math edu site.
lol.
MO is brutal as far as killing off marginal/borderline stuff, even more so than many se's.
its an extraordinary case that shows there can be a huge amt of experts and very high constraints on quality.
 
As well they should (given the existence of math.se)
same for tcs
 
vzn
6:08 PM
actually math.se and cs.se arose after the elite sites MO/tcs.se were created 1st. but yeah.
 
I think that in this case the "quality questions" thing is releavnat
 
vzn
yes, basically agree.
 
Yes, I know.
It's actually backwards.
 
vzn
have asked research questions on tcs.se and still gotten heavily roughed up over them.
 
MO is the core, and math is the "overflow"
 
vzn
6:09 PM
its not just research but elite research sometimes.
←always a bit terrified to ask questions on tcs.se/MO even after yrs of doing so.
 
I've never asked a question on either.
 
vzn
so yeah, have a huge luv/hate relationship with se myself.
some days luv/hate other days hate/luv :p
 
vzn
the software capabilities/features are amazing/awesome at times. the use of them (by humans) is whats often killer
have you heard about the new meta stackexchange reorg?
you can try raising some issues there.
bzzzt, zap j/k :p
 
Yes, I saw the meta reorg.
 
vzn
6:16 PM
Tim Post on April 17, 2014

We’re very pleased to announce that as of today, we’re (finally) splitting the site formerly known as Meta Stack Overflow into two sites:

Meta Stack Overflow is a brand-spanking new site for discussions specific to the Stack Overflow (programming) community:

Meta Stack Exchange will cover feature requests, bugs, and any discussion topics that affect the entire network:

Meta has always been one of the most important things that make our network… work. Almost all of our most important features, improvements, and community rules were partly or entirely based on our users telling us how we could help …

whole new playground
with brand new bullies :p
meet the new boss, same as the old boss =(
 
Speaking of "boss"...
The phone meeting that I looked like I was attending just ended.
 
vzn
hah ok
 
So I had better actually do some work now. :p
 
vzn
me2 nice chat l8r smart dude
 
you too. thx!
 
vzn
6:22 PM
=D
a zen question to ponder, if undergrads arent allowed to ask about hw on this site (the main audience) then how can they use it, how do we draw em in, make it "sticky"?
 
 
4 hours later…
vzn
9:56 PM
alas, se is in many ways the "only game in town right now". :(
 

« first day (771 days earlier)      last day (3674 days later) »