« first day (29 days earlier)      last day (4445 days later) » 
00:00 - 14:0014:00 - 00:00

2:05 PM
@MrWizard No, it just works in Chrome. I even see a list of all users starting with the name.
@MrWizard look here
 
@halirutan I see the same list but from what I can tell I do not have auto-complete.
 
Not with Tab?
 
@MrWizard Time to put my ex-MSO junkie skills to good use... Here's some suggested reading: 1, 2, 3, 4 :)
 
2
Q: Is posting an incomplete answer then immediately improving it helpful or aggravating?

Mr.WizardRecently I have been posting an answer that is not yet complete and then continuing to make additions over the next few minutes. Undeniably I enjoy getting first post, but there is another reason that I have been doing this: to get some ink on the page so that others know what is coming. I pers...

 
Hey @yoda... I need a favor from you. Got time?
 
2:24 PM
@MrWizard I'm always amazed how many type and language errors I make which I don't see. Even after the second time of reading. Thx for correcting.
 
@halirutan I've been making a lot this morning myself. A bit fatigued I am.
@yoda thank you I am reading now
 
...I presume you guys know it's easier to produce buggy solutions when you're sleep-deprived, no?
 
@JM what's up
 
@yoda Does that computer you're using now have MATLAB?
 
yeah
 
2:31 PM
If it's not too much trouble, could you execute colormap jet; cm = colormap and give me the matrix produced as a plaintext file?
(Should be a 64-by-3 matrix, if memory serves correctly.)
 
@JM heh, way ahead of you (by a few months) :D
5
Q: Custom ColorFunction/ColorData in ArrayPlot (and similar functions)

yodaThis is related to Simon's question on changing default ColorData in Mathematica. While the solutions all addressed the issue of changing ColorData in line plots, I didn't quite find the discussion helpful in changing the ColorFunction/ColorData in ContourPlot/ArrayPlot/Plot3D, etc. TLDR: I...

there should be a text file and a download link
 
@yoda Aha. Thanks! I already managed to fake all the other MATLAB colormaps in Mathematica except jet... :)
 
btw, you can directly do cm=jet; :)
unless of course, if you have a variable named jet in the workspace
 
Uh-oh, I forgot that Megaupload is kaput now...
 
did you check the pastie? (2-3 lines under the megaupload link)
 
2:37 PM
Okay, so my eyesight isn't as good as it was... thanks!
 
you're welcome :)
 
Actually, I already had printouts of the M-files in hand, but jet was troublesome to translate in Mathematica...
 
Have you also exported mma's ColorData as a colormap for use in MATLAB? Or is that suggestion blasphemous?
 
speaking of MATLAB vs MMA -- matlab has a nifty "realtime zoom" tool for plots
where you can zoom in and gain better resolution in a little region
then zoom back out to the whole thing
know of any tricks for something like that in MMA?
 
@yoda That might be a good idea too... I like "TemperatureMap" and "DeepSeaColors"...
 
2:44 PM
Yeah, some of them are pretty neat. I usually use my own maps depending on what's being plotted, so haven't done so yet
@EliLansey hmm... I thought I saw an SO question on that (I might be wrong)... let me look
 
Actually, after some analysis, apparently rcollyer's color scheme from a while back and MATLAB's hot are eerily similar.
The positions of colors are slightly different, but the effect's the same.
 
@yoda found it here
 
@JM I didn't know he created that himself... I just assumed it was the hot colormap :)
 
sort of
 
@EliLansey Although not the same, take a look at Experimental`Explore[]
 
2:50 PM
@yoda Nah, I compared RGBColor[]s. Stuff's shifted a bit differently...
 
@yoda wow, that's pretty nifty
 
What's the difference between BeginPackage["`name`"] and BeginPackage["name`"] (note the extra backtick at the beginning of "name" in the first one)?
 
@yoda That's for inheriting contexts, IIRC. Let me check my notes...
@yoda Ah. Try this: BeginPackage["`test`"]; Print[Context[]]; EndPackage[]; and BeginPackage["test`"]; Print[Context[]]; EndPackage[];
 
@JM thanks, that makes it clear :)
 
What is the point in writing some symbol names as `x then (in a package)?
 
3:03 PM
@JM I think that may be where I got the inspiration. I honestly don't recall.
 
@Szabolcs That's often to ensure they're private symbols... IIRC.
 
@JM It's shifted to limit the low level noise from showing up.
 
@rcollyer Don't worry, it's not an exact copy. I checked. ;)
 
@JM I mean, in what situation will x and `x be different?
 
@JM I wasn't worried. Different implementations and all, so I'm covered.
 
3:05 PM
OK, got it. It forces the symbol to be part of the current context
 
@Szabolcs I can see where that would be useful to guarantee that no shadowing occurs.
 
x
Begin["zoo`"]
Context[x]
Context[`x]
End[]
 
@Szabolcs Yeah, I was about to type that after checking my demo... :)
 
Suddenly the difference between Begin[] and BeginPackage[] is a lot more clear.
BTW I cannot type x as inline code here. It won't render: `` x``
 
In what way, it alludes me at the moment.
 
3:09 PM
@Szabolcs Yeah, if you'll note the structure of a package, you have BeginPackage[], usage messages, Begin[] (private context, natch), definitions, and then End[] and EndPackage[]...
`x
x
Hmm...
Nah, I use backslashes.
 
Ah, escaping them.
 
`\`x` gives `x.
 
:P
 
There. :P
 
@JM Thanks
 
acl
3:13 PM
@EliLansey or this?
16
Q: dynamic interactivity problem

aclI am trying to have two panels, the left showing a graphic and two locators, the right one a zoomed-in version in the area defined by the locators. I've tried ClearAll[mndpt]; mndpt = Compile[{{c, _Complex}, {maxiter, _Integer}}, Module[{z, iters}, iters = 0.; z = c; ...

 
@acl also, sort of. basically, it seems like in MMA we need a cludgy workaround. shame. that's one place where matlab seems better :(
also, also, how do you paste the links in that nifty formatted form in chat?
 
just paste the link alone.
 
Yeah, the bare link suffices.
(For the record though, I'm not terribly fond of it...)
 
I think I am addicted to this site and it is harming my work. I need to try not to open it for a while.
 
3:20 PM
@Szabolcs yeah, ditto.
 
The first step is admitting you have a problem... :)
2
"Hello, I am Szabolcs, and I am an addict."
 
So if you don't see me for a while, you know what happened
 
You could make it just every two days, or something.
 
So, you're joining MMA: MatheMatica Anonymous?
 
@rcollyer version 8?
 
3:23 PM
Or later ...
 
8.0.4?
 
Possibly. Or, 9 if it is coming out soon.
 
What does "soon" mean? :)
 
In the same sense as Duke Nuke'm Forever.
3
 
Why did I know you'd give that answer... :)
 
3:26 PM
Why do I know what @rcollyer is referring to?
 
@JM Ow. just because I'm nerd, does not mean you have to point it out. :P
@Heike That game was supposed to come out 10 - 12 years ago. It came out last year, and it was horrible.
 
@rcollyer Nothing of that sort. If we switched places, I'd have answered in the same way... :)
 
@Heike To all reports. I haven't actually played it.
@JM Ah, point.
It's weird. I've pegged the rep cap so hard today that I'm six votes over the limit.
 
@rcollyer Yes I know. I was just wondering why my head is filled with pointless trivia like that.
I guess I spent too much time on the internet
 
I don't ask those types of questions, I may not like the answer.
See you all later.
 
3:34 PM
@rcollyer It's a pity that extra votes don't spill over to the next day.
 
I understand the reasoning though... similar to why Mathematica has things like $RecursionLimit...
 
@JM I understand they put a cap on the amount of rep you can get, but maybe the limit should be a function of the number of answered/asked questions as well.
 
@MrWizard Congrats.
 
@Heike That's true.
 
3:40 PM
@MrWizard Nice. :)
 
@MrWizard Congrats. That's very nice.
 
@MrWizard Congrats :)
 
I wonder what my counter for upvoted competing answers is.
 
@Heike You've answered 52 questions, so it may be up there.
 
Thanks. While I have your attention:
33
Q: Gold Sportsmanship badge

Mr.WizardI propose a new Gold level Sportsmanship badge. I earned the silver badge June 4th, and I continue to vote for competing answers. I would like to feel that I am working toward something, not because I would otherwise cease to vote for these answers, which clearly I have not, but because it adds...

 
3:42 PM
@Heike Mine is no where close. I only have 20 answers.
 
@rcollyer it's not by # of answers, but # of competing answers... I think I should be there soon
I don't have enough answers either
 
@yoda are you sure?
 
@yoda But you can only vote for competing answers if you have answered the question yourself
 
@yoda I know that, but I'm limited in opportunity to upvote competing answers.
 
That is do you need 100 of your own answers or not? I have assumed that it only counts one for each of your own answers. Otherwise this would be fairly easy to game by posting to one question with 100 answers and voting for all of them.
 
3:45 PM
 
If you find a Meta question let me know please. That sounds like something to be fixed.
 
@rcollyer I know, me too... which is why I think I should be there in a few more days (perhaps after I've answered 60ish answers)
@MrWizard why should it be fixed?
@Heike that's correct... some questions have more than one competing answer, so upvoting 2 of them counts as 2, not 1.
 
@yoda While I admit to my addiction, I do like the pace at which I answer things. Unfortunately, that means I have to wait for some of the cooler badges.
 
@rcollyer yeah, same here. Sometimes, it's a bit disappointing that ones I could've answered easily are already answered and the ones I can answer, but need time are answered by someone more skilled... but I'm not going all out (not that I'd achieve much if I did...)
 
Only 13 more votes and I'll have gold on meta.so!
 
3:49 PM
I need double — 26 to reach mine =)
 
@yoda Exactly. I try to nab the ones I can answer reasonably, and if someone beats me to it. Oh well. At that point, I'll answer only if I can add something.
@yoda it's the gift that keeps on giving. I post so rarely on meta, this is funny.
 
@yoda bonus.
and, yes I already upvoted it.
 
@yoda let me see if I understand this. Post an answer to this question, get at least one vote on it, then upvote all the other answers and POP, sportsmanship badge. That doesn't seem broken to you? Or do I misunderstand?
 
Did anyone notice, they've started posting answer stats on the tag pages?
 
3:53 PM
No, that's new to me
 
@MrWizard again, me to the rescue!
CW posts don't count... and any post that has more than 30 answers automatically becomes CW
@rcollyer yeah, they're undertaking a major tag-page, tag-wiki overhaul
 
@rcollyer I've noticed that.
 
(me thinks yoda is about as obsessive of Meta as I am of this site...)
 
@MrWizard used to be... moderating 2 sites kind of sucks you into it, which is why I was warning you folks to not fall into the same trap/burn out
 
@yoda I figured that by Sam Saffron's comments.
Our stats on SO are pretty bad (for us) for the past week/month.
I wonder why that is ...
 
3:56 PM
@rcollyer My stats would be better if they didn't keep migrating questions.
 
@Heike that's true. But, I keep flagging for migration ...
For this question, is a better choice than ?
 
Why is there a path tag?
 
@rcollyer You bastard -- how am I ever going to reach trusted user now? (tongue firmly in cheek)
 
@Heike only on that question.
@MrWizard Hit 4k! :P
 
I mean SO ;-p
 
3:59 PM
Who cares about SO? We have our own playground now.
 
I think is a bit more transparent a tag...
 
That's what I thought. Editing ...
 
I wonder when there'll be a MATLAB analog of this site...
 
I wonder what my non-mathematica-tag votes total; not much I think.
 
@MrWizard I still need to earn access to the moderator tools.
 
4:01 PM
@JM possibly never.
Feb 11 at 18:38, by yoda
@Heike just share a link to "How can I find waldo" on reddit again... those folks there have a short memory
It's pretty simple... you just won't do it =)
 
Oh CRAP.
 
one of us should do it. Anyone with a moderately active account there?
 
@MrWizard mine isn't to bad outside of mma, mostly do to the boost libraries.
 
hm... damn clumsy ass that I am look what I did.
 
@yoda You're right about that one.
 
4:03 PM
@MrWizard ???
 
Why am I replying to my own comment?
 
I edited Heike's comment instead of replying to it.
 
At least you noticed what you did. Could be worse.
 
Bleh, it seems I am getting these "The notebook's file has changed outside of Mathematica..." errors more often.
 
@Heike I was trying to say I'll see if I can help with that. Bleh.
 
4:04 PM
I wonder what the most lucrative tags on SO are
@MrWizard Better not, or you might mess things up ;-)
 
<:-o
 
:P
 
@MrWizard lol... you're a mod on all of chat.se :) can happen in any chat room
 
Seriously gotta get something done for tomorrow. See you all later.
 
bye
 
4:06 PM
@MrWizard promise to kick me out if I show up here for more than 5 mins today.
 
@rcollyer if he won't, I will... starting now!
 
lol --- if I'm around :-)
 
@yoda bye.
 
bye bye... kicks
lol. that was fun =) We need more willing folks :D
 
too late
@yoda did you actually kick him off?
 
4:08 PM
@Heike yes
 
Don't let the power go to your head.
 
I have to leave now. Later, y'all.
 
acl
Am I saying something stupid in my comment here?
0
Q: Where can I permanently modify $Path?

murrayI've quite forgotten: what file does one modify in order to add directories to $Path globally? Specifically, I want to include $UserBaseDirectory/ExtraPackges. Or does one do it by means of modifying some setting in the OptionInspector (for Global Preferences) and, if so, which setting? I didn't...

or is it really that easy?
 
@acl I think that should work. I haven't looked at it but I would have thought that there is some option in the Options Inspector that sets the search path.
 
4:23 PM
1
Q: Should we migrate the "Toolbag" question?

belisariusOur dearest SO "ToolBag" question has been closed. Should we migrate it here for re-openning and caressing?

 
acl
@Heike I was trying to work out if he really is asking this. I generally prefer adding stuff to init.m rather than options (since then I can just look at init.m and see what's happening)
 
@acl It looks like $Path is set in $InstallationDirectory/SystemFiles/Kernel/SystemResources/MacOSX-x86-64/sysinit.m (for OS X at least)
it's as simple as $Path={list of dirs}, so I don't see why you can't just add to this list in init.m
 
acl
@Heike yes I have posted an answer along those lines. if you can work out how to do it in the option inspector go ahead and add that (I get confused going in there)
 
4:40 PM
Lately simple questions get no or very few upvotes. I think this is a good (though basic) question: mathematica.stackexchange.com/questions/1874/… It is not that easy to figure this out on one's own.
 
@Szabolcs Maybe people are only upvoting questions they find interesting or challenging?
 
True.
 
acl
haha, someone upvoted my answer and then removed the upvote :)
 
@acl At least that's not as bad as someone accepting your answer and then unaccepting it
 
4:57 PM
@Szabolcs I think that parallel addition of matrix question is a person coming from MATLAB where it's faster to first initialize the table, then fill it.
 
@Heike haha, nah... I've never used it before, and I've had it for abt 8-9 months... @rcollyer was a good lamb =)
 
@EliLansey But he said that his real problem is much more complex than a matrix addition ... and there's something weird happening, look:
In[1]:= LaunchKernels[]
n = 1000;

Out[1]= {KernelObject[1, "local"], KernelObject[2, "local"]}

In[3]:= a = RandomReal[{}, {n, n}];
b = RandomReal[{}, {n, n}];

In[5]:= MapThread[Plus, {a, b}]; // AbsoluteTiming

Out[5]= {0.0270015, Null}

In[6]:= Parallelize[MapThread[Plus, {a, b}]]; // AbsoluteTiming

Out[6]= {0.3650209, Null}
If I use Method -> "CoarsestGrained", it is even slower!
 
@Szabolcs I generally avoid using Parallelize whenever possible, since it's usually not as smart as a human
 
@EliLansey But it can parallelize things like MapThread, Outer, etc. as well, which is a big advantage ... ok, let me try ParallelTable
 
@Szabolcs That's true
 
5:01 PM
Much better now:
In[9]:= ParallelTable[a[[i]] + b[[i]], {i, Length[a]}]; // AbsoluteTiming

Out[9]= {0.1240071, Null}
 
@Szabolcs Yeah, see my solution
 
This should be equivalent to the MapThread above.
 
0
A: Efficiency of ParallelDo when a global variable./matrix is being updated

Eli LanseyIs there any reason you can't directly construct the Table, as in: DistributeDefinitions[a,b] c1 = ParallelTable[b[[i, j]] + a[[i, j]], {j, 1, n}, {i, 1, n}] On my machine, with n=500;, Table[b[[i, j]] + a[[i, j]], {j, 1, n}, {i, 1, n}];// AbsoluteTiming takes 0.59 s, while ParallelTable takes...

for n=1000 plain-old table takes over 2 secs
paralletable under .5
 
For n=5000, ParallelTable runs in 3.5 seconds here while Table runs in 0.3 ...
But I only use a single iterator in Table, not two
@EliLansey How many cores do you have?
 
5:09 PM
8 real ones (i.e. not counting hyperthreading or anything similar)? I wish I had that :-)
 
@yoda I was actually hitting the leave button. So, if you kicked me off, or I kicked myself off is up to the logs ... :P But, thanks. I'll need it today.
 
@Szabolcs 8 real ones :)
we have some here hyperthreaded to 16
and that's just our workstations
i have 8gig ram, which is on the low side
usually we're 16-32
and then we have a cluster ;)
 
@rcollyer I fully understand that... I was pretty squeezed during my qualifying last month and that was right in the middle of private beta!
Well, I didn't log out and it went ok, but I would've appreciated someone kicking me out. You probably need a lot more kicking now =)
 
@yoda I have a presentation tomorrow. So, probably will need the boot on occasion.
 
@Szabolcs for n=500 Table takes ~55s, ParallelTable takes ~11s
n=5000 rather
and the full 2D table
 
5:14 PM
@rcollyer Have you finished the slides?
 
Table[b[[i, j]] + a[[i, j]], {j, 1, n}, {i, 1, n}];
 
@rcollyer is this thesis related or the alma mater stuff you were talking about?
 
@EliLansey Use Method -> "CoarsestGrained" and launch 16 kernels if you have hyperthreading and enough kernel licenses :-) I wonder how much it'll cut from the time
 
@Heike no, and that's the problem. I could do it all on the black board, I'd be fine. But, I really can't.
@yoda for the alma mater.
 
made it sound like dark matter
 
5:15 PM
@Szabolcs on the 8, coursestgrained takes longer, weird
 
@yoda alma matter: where students orbit indefinitely hoping to graduate one day.
Anyway. Time to reorganize some slides.
 
acl
@Heike ah yes
 
I have to go as well
 
bye
It seems like I'm the only impolite ass who just leaves the computer (while logged in) and goes to bed without notifying others that I won't be back to reply for the night...
 
5:56 PM
@halirutan Can you check the highlighting in this answer? Why is "Warning" green in the message? Did DL copy it incorrectly?
 
@yoda It's normal. It's because of the : character. It's interpreted as a named pattern.
It's not code, just a message text, but the highlighter has been dealing with message texts pretty well so far
 
@Szabolcs Yeah, I guess my question was (if I understood the script correctly), why doesn't it highlight everything after :: as message text?
 
@yoda how would you highlight the following then:
fun::usage = "fun[x,y,z] text text text";
The next question would be: why not highlight everything as text after something like function::message: (i.e. trailing colon). That would be nice, but what if the message text is multi-line?
 
6:14 PM
@Szabolcs Hmm... good point. I guess it's alright.. This is the first time I've come across the script not highlighting (you've probably found them all before I had a chance to see :D), but I don't think it's worth halirutan's time to fix this small issue
 
Not only that, but I'm not even sure how it could be fixed while making sure that no real-code highlighting is broken ...
 
@Szabolcs How do you read the code (or where do you find the m file) for mma supplied packages like Combinatorica, for instance
 
installDir/AddOns
Packages is old stuff, Applications is new stuff, the rest is mostly not well documented or very old stuff
Combinatorica.m has no comments whatsoever ...
 
6:32 PM
@Szabolcs perfect. I guess I was looking in the wrong place.
@Szabolcs Yeah... they probably want you to buy the book :)
 
It's in a really messy state now, with all the conflicts with the builtin Graph. But I find the builtin Graph functionality messy, incomplete, and it has silent bugs which make you end up with wrong results
I am really really really hoping for a significant improvement in 9
 
Yeah, I'm just trying to pick up some package writing hints as I read along (and someone mentioned that a lot of packages are full of goodies)... a lot of my code is just in monolith notebooks as i hack and patch together stuff to get work done and I need to fix that
 
Out of random curiosity, anyone here attending the APS march meeting in ~2 weeks?
 
nope
 
@EliLansey No, my stuff wasn't done in time, and won't be done in time to make my thesis. Oh well.
Unless, I can pull a miracle out of unusual places.
But, either way, I won't be in Boston in 2 wks.
 
6:45 PM
@rcollyer ok, fair 'nuff
 
@rcollyer hey, stop reading the transcript and dropping in to comment!
 
@yoda You got the order wrong. I drop in first, then read the transcript, then I comment, then I go.
 
now go!
 
Bye.
 
bye
 
7:33 PM
The tracking stats are now live on Area51. We're mostly green (good) with two orange (okay). So, as far as these stats are worth anything, we're doing good!
 
7:44 PM
@yoda Unfortunately, people copy such messages too slopy as Mathematica-input ;-)
Since it is no valid syntax, the pattern-matcher assumes it is a M-pattern. Even in Mathematica this is not highlighted correctly.
I was thinking about this problem earlier, but since those messages can go over several lines and contain really everything, from keyword to numbers and since they are not in quotes like strings, I see no easy way to match this.
When I include messages in a post, I surround them by comment-braces.
 
@Verbeia The leveling off of traffic is a good thing. It means we are firmly within the period they use for the moving average, and an uptick in that value will mean we are showing signs of heading towards graduation.
 
8:08 PM
Okay some irritation here. I have a function f[x] that I can evaluate to two decimal places via N[f[2],2]. Great. Now if i do either of the following N[f[2]*<float>,2] or N[N[f[2],2]*<float>,2] Mathematica ignores my instructions of N[ ,2]. What am I doing wrong?
 
Hint N[f[2]*1.0`10, 2]
 
What is ``10`?
 
@LordStryker ahh, come on. Don't you have a Doc-Center?
 
Well here's a potential issue, my <float> is stored in a variable that I'm calling.
so would it be N[f[2]*var10, 2]`?
@Halirutan, Myself and my colleague have been beating our heads over such a simple issue for the past 20 minutes. The program is not doing what the documentation is seemingly implying though we're obviously missing something.
 
Give me a sec. I'm eating ;-)
 
8:18 PM
No worries. We'll figure it out eventually.
 
@LordStryker Ok, do you know what happens when you do N[2,2]?
 
yes
Okay, so one decimal place. Typo in my original message.
 
If you set the precision explicitely, then you loose information which cannot be recovered.
The thing with N[2*var,1] is that var has $MachinePrecision
 
@halirutan For the evaluation of my f[x]*var, I need the output to provide me with 3 numbers (giving me one decimal place).
 
If you multiply them, the result has $MachinePrecision too.
 
8:26 PM
So how can I circumvent this?
 
And as stated in the doc for N: With machine-precision input, the results are always machine precision.
 
Okay
thanks.
 
So question, do you want to print your number or do you want to reduce your precision?
 
I would like to keep the precision in the calc, but then print to the screen a rounded off number
 
NumberForm[var, {1, 2}]
@LordStryker Then you don't want N as I assumed in the first place;-)
 
8:31 PM
@halirutan This worked. Thank you.
@halirutan You assumed correctly. Gold star for you.
 
@LordStryker You're welcome.
 
 
3 hours later…
11:08 PM
@Szabolcs @Szabolcs for Mathematica (not counting things like Dot, LinearSolve, etc., at machine precision that are treated using MKL), hyperthreading cores should be nearly as good as real cores. (Admittedly I don't have a computer with hyperthreading here to check that. But, I feel it is likely to be the case since Mathematica code generally achieves low IPC due to low arithmetic intensity, large working set, and complex memory access patterns.)
 
@OleksandrR You are quite familiar with the PCT, right? Maybe you can answer here: mathematica.stackexchange.com/questions/1883/…
 
Not sure I've got anything to add that you can't read in the docs or in the source though. The existing answers are pretty good.
 
@OleksandrR In what time zone are you? GMT+3? Here's it's 1 AM, I'm going to sleep now. Good night!
 
On the other hand I also thought that everything there was to say about the compiler had already been said lately, and Leonid proved me wrong by posting a nice overview. So maybe someone else will chip in.
OK, goodnight! (GMT here.)
 
11:24 PM
@J.M we were editing at the same time
You were just 5 secs earlier ;-)
 
00:00 - 14:0014:00 - 00:00

« first day (29 days earlier)      last day (4445 days later) »