« first day (29 days earlier)      last day (3626 days later) » 
00:00 - 14:0014:00 - 00:00

12:52 AM
@All, I hate Excel. It does nasty things to perfectly good data. And, yes, Excel was my tool of choice for years.
Well, old-school Excel wasn't so bad. The current version is a monstrosity that should be shot behind the barn.
@rcollyer I thought everybody hated excel?
(I've only used it briefly and it seems powerful to be honest)
well, "used", I've played with it
It *was* good... somewhat.
1:09 AM
@acl Leonid's answers are always very informative.
@rcollyer Excel does a decent job at being okay. :P
I used to generate graphs in Excel, but I use another program now. I still use Excel to create my tables, that I then convert to LATEX. I find it too laggy for any other use.
If anybody's interested, the graphing tool I use is called Plot. It is an OS X app, available here
Eep, Apple...
@JM why?
'cause I can't afford those computers... :D
Ah. I thought there would be an ideological argument offered.
@CHM I do have one, but this chatroom isn't the place for holy wars like that... ;)
1:18 AM
Also my suit of armor shrunk in the wash.
@CHM For me, the last straw with Excel was it stripping off the zeroes beyond the decimal points, reducing the precision of my data.
@rcollyer Nasty...
Don't get me wrong, I've done some amazing things with Excel. I even saw an entire paper path simulated in Excel, complete with animation.
@JM Apple's customer service is what I prefer about the company.
1:22 AM
But, it was way to easy to cross the line into territory where it couldn't handle what I wanted it to do.
@rcollyer gneh..
@CHM Oh, I've heard they are indeed usually nice...
@JM I'm typing this on a mac. :)
@rcollyer Well, I've always been of the opinion is that it takes a bit of effort to adapt Excel to technical uses...
@JM I bought this laptop in 2010, with a 300$ AppleCare extended warranty.
1:23 AM
@rcollyer Heh. I saw the transcript BTW; I like the shadowy effect. Mind if I borrow your coloring scheme for something I'm doing?
@JM true, but you have to delve into VBA very, very quickly.
@JM go for it.
Just before Christmas, my HDD failed, a bunch of pixels on my screen discolored, and a couple of keys on my keyboard went loose.
Went to the Apple Store, they asked for my name and birthdate, took my computer, replaced the parts, and I had my laptop back - in one evening.
1000$ worth of reparations. Now.. yesterday my battery gave me a "Service Battery" alert. That's not good news though, because the battery I have to pay.
But meh, to me it's worth it, not having to worry about customer care.
@rcollyer It looks good on my artwork so far. Thanks!
@JM I want to see the pics.
@Both do you do Mathematica artwork?
1:29 AM
Personally, no, but I like my visualizations to have a sense of style.
@rcollyer Maybe later; I'm still figuring out this Perlin noise business...
...the experiments are encouraging, though.
@JM ooh. that would look great with my "hellfire" color scheme.
@CHM I've been featuring my artwork over a series of Gravatars.
and, yes that was along the lines of what I was thinking when I put it together.
@rcollyer That's exactly why I got excited when I saw your color function... :)
1:31 AM
It is a shame, the details don't show up well in my gravitar.
(BTW: my last blog entry was on Perlin noise experiments.)
@rcollyer Yeah; that's why I've taken to showing slightly bigger versions of my Gravatar on my profile for some of my artwork.
@JM where can I see your gravatars in their original format?
@CHM I haven't thought of where I could post a gallery. Flickr and DeviantArt don't sit well with me, for some reason.
@JM good point.
@rcollyer is your gravatar a cellular automata?
1:40 AM
@CHM No, it's a spectrogram of a sound file off of my pc (I think, it's been five years).
Good point @JM
Although I don't really like @@@
Implemented it anyway. Now I feel like a hacker
@David That's okay; I wanted to mention it since a number of people prefer that...
Oh, I just realized I wrote pretty much the same thing as István Zachar.
I'll leave the answer there anyway because he doesn't explain anything.
(Bonus points for me reading his response, not realizing that it's my exact idea)
@JM By the way, how did you do the honeycomb torus?
Is it physical or just texture?
@David I used a Perlin noise Texture[].
Wait, did you mean my current Gravatar, or the hexagonal tori I had previously?
@rcollyer ok.
Q: Can Eigenvalues[] and Eigenvectors[] be assumed to return the same ordering?

Peeter JootIf I do back to back calls of Eigenvalues[] and Eigenvectors[] can these be assumed to order the values and vectors the same, or is each independent? Related to this is a curiousity about the performance of a pair of calls like this. Are mathematica optimizations done at an expression level, or...

I'm hesitating to post a comment. I don't quite know what he wants.
1:47 AM
@CHM I left one; if he wants both, there's already a function that guarantees correspondence.
My impression is, he could try it and see. But then, I don't quite understand his question.
@CHM I posted an answer.
I saw. Maybe I'll understand his question better by reading the comments/answers :P
Hmm. Ok.
You could tell him how to Orthogonalize[] and Normalize[] his vectors too.
@CHM I will when I got off the phone.
1:51 AM
@JM Your ex-gravatar.
@rcollyer Added a comment.
@JM Wrote an answer unrelated to you guys while you added the comment.
@David Ah, that was a rewrite of Mark's old code.
I think it's "David reinvents the wheel" sunday.
Wait, it's not sunday
Geez. Something's wrong.
@David I basically generalized and updated it for version eight.
1:53 AM
"Mark's old code"?
Is that a thing?
Should I find it somewhere on his website?
@David Apparently it's not there anymore, so I had to check the Wayback machine. See this.
The machine that serves this file is down. We're working on it.
Oh, the older version works.
Apparently the SE engine munges the colon. Change the "%3A" as needed.
In any event: I still have to tweak the code further to make it user-friendly. As it stands, I'm the only guy who can read my code... :D
I know that feeling.
2:03 AM
@JM what?
Reading my updated answer?
Oh, visitors per day is declining.
Sorry, my computer froze mid-upload...
@JM Awesome!
@rcollyer I know a good color scheme when I see one... :) You see why my excitement was justified?
@rcollyer Oops, didn't see your edit until I posted.
2:08 AM
@rcollyer Too bad I already upvoted. It's a bit annoying that OP neglected to mention what kind of matrices is he dealing with...
@JM ten to one, Hermitian. He does a lot of physics stuff.
@rcollyer Then he won't have to worry about defective matrices...
True, but it was a nice add.
+1 for spectral decomposition :P
Agreed, it's nice to be general.
2:10 AM
Oh, I added my gravitar to my profile.
@CHM thanks.
@CHM If we must be precise: SchurDecomposition[] is what actually returns the "spectral decomposition". Eigensystem[] returns what it says. The deep theorem is the correspondence of the outputs of those two functions...
@both have you noticed how Mathematica orders the eigensystem, depending on whether you're using symbols or Integers in your matrix?
It's no big deal, but I don't know why it would do that.
@CHM Yeah, the ordering can be spotty. That's why I always tend to use only Eigenvalues[] or Eigensystem[].
@JM Never heard of SchurDecomposition before.
@JM I was not aware. I may have to update the post, then.
2:13 AM
I haven't had a situation where I only needed eigenvectors.
@JM good point, neither have I. Which is saying something considering the field I'm in.
@rcollyer Well, SchurDecomposition[] works only on inexact matrices. So that's one caveat for demonstrating the theorem...
Define "inexact."
@rcollyer the output of N[stuff, prec]...
Q: How to plot layered graphs?

István ZacharGiven a directed acyclic graph, there is always a possibility to plot the graph as layers, where nodes always send edges into one direction (usually down), to successive layers, and never backwards. The "LayeredDrawing" option value for GraphLayout fails miserably. There is the built-in (but some...

2:15 AM
@JM point.
wouldn't cah's comment qualify as an answer?
It would be nice if SchurDecomposition[] can be made to work on exact matrices, but apparently they haven't done that yet.
...with that, I need to be away for a while. Later!
See you, later.
I'm off too, stuff to read.
2:30 AM
1 hour later…
3:31 AM
26 questions in the last 24 hours!
Here's a better version of my gravatar:
via ColorSeparate[Image[...]][[1]] // ColorNegate.
@Verbeia We're slowly moving up the rankings. Currently, we appear to be in a dead heat with DSP. With luck, the influx of questions will mean we'll stay ahead of them.
4:19 AM
First, Taxonomist Badge was awarded for . will be there soon.
@rcollyer what's DSP?
@rcollyer nvm.
I think we're doing a lot better than them, actually.
We are. It just took us time to for the averages to work out better than theirs.
We need at least twice as much visitors per day.
Why did this get downvoted? Did I miss something?
In fact, earlier today, we were ahead of them, then we weren't for a good part of the day. Now, we're ahead of them, again.
4:27 AM
@yoda I missed it too, then.
@yoda I don't know.
The guy wants to delete his question, I guess some people helped him getting no upvotes...
@yoda do you know why this question shows no closing info?
Ah, I see. It was in the middle of being merged.
@yoda, I don't think that question should have been closed personally.
@rcollyer, hello
4:43 AM
Hola. I was about to go to bed when I saw you in here.
@rcollyer, its been a good last couple of days on here for me :)
@AndyRoss yes it has. Now I have two WRI employees to overcome.
@rcollyer, of course my productivity here is inversely proportional to my productivity at work :)
Yes, my goal is to just once be ahead of all WRI employees here.
@AndyRoss don't I know it.
I think you may have your work cut out for you.
4:45 AM
Speaking of which, have you made any progress on your BF interpreter?
Oh, I sort of put it to bed. I had some fun with it, got it working and then archived it.
There are only two acknowledged employees ahead of me.
Nevermind, I have moderator capabilities now :)
Almost for me.
I'll check to see if there are any of your questions/answers I've neglected to vote for. Just a sec.
Well, that should help :)
4:51 AM
That was a little weird watching my rep go up like that.
@rcollyer: aw, I liked the burning coals version...
@JM yeah, but it wasn't all that visible.
Do you have any other nice color schemes? I'm really digging this...
@rcollyer, I think that should do it. I thought I kept up with voting better than that.
@yoda, you there?
4:53 AM
@AndyRoss cool. Thanks.
@JM i'm looking for a plot, give me a sec. It isn't as cool, though.
@rcollyer, you wrote the +1 worthy Q&A I just voted :)
@JM yeah
@rcollyer (BTW: this one isn't Perlin noise anymore; it's Fourier-transformed attenuated noise.)
@AndyRoss which one was that?
@JM and the difference?
@yoda Did you get the bit about the image processing question?
4:55 AM
@rcollyer ?
@rcollyer Perlin doesn't need Fourier transforms... just Hermite interpolations.
@JM Yeah, I think I also replied to you
We'll take that question.. .it'll be a good fit
@AndyRoss sorry, I did not understand your prior comment: "you wrote the +1 worthy Q&A I just voted."
@rcollyer, meaning no need to thank me, I just gave upvotes for things that deserved upvotes.
I flagged it too, but in my experience, the math mods are extremely slow... sometimes, flags sit for a day or more
4:57 AM
@AndyRoss Now, I understand.
@yoda Every single one of them is busy, busy, busy... :)
WooHoo! just got access to moderator tools!
@yoda so high-rep users tend to pick up much of the slack there these days.
@rcollyer and now your productivity really starts to slide :)
@rcollyer prepare to be let down =)
4:59 AM
@AndyRoss what productivity. I have a presentation on Friday that I have to finish tomorrow.
Yeah, the novelty of the mod panel wears off after an hour or so... :P
@rcollyer bleck. I'm thankful that I only have to do a couple of those a year. The big ones being the user conference.
I just like viewing deleted answers.
@AndyRoss its for my alma mater, so I'm looking forward to it.
not that it's completed, yet.
@rcollyer, did one of those myself not too long ago.
It's to undergrads, so I'm trying to give them a sense of computational physics without going into any real details.
Because details would make them run away screaming.
5:01 AM
@rcollyer, likely they would make me run away screaming as well :)
@yoda: after you told about your having to bitch at the math mods, I've become more careful about flagging stuff for migration...
@AndyRoss some of it is rather interesting. most not.
Unfortunately, the need for computation occurs when you add the second particle in quantum mechanics.
@rcollyer: the current color scheme in your Gravatar is just GrayLevel[]?
@JM heh, thanks :)
@rcollyer my background is Biology and Stats so you've already lost me just by mentioning quantum mechanics :)
5:03 AM
I've been really slack on dsp... questions and views are zooming up, but no one's picking up the slack in editing and tidying... normally, I'd do that routinely, but got bored and burnt out on gardening
@yoda Now you can't claim I don't give a damn... :P
That's one good thing about mma.se community... everyone chips in for everything
@JM ColorSeparate[Image[ ... ]][[1]] // ColorNegate as I couldn't regenerate it.
@rcollyer Ah, much simpler than what I was thinking...
@AndyRoss ah, high level quantum, I mean physics of valence shell electrons, I mean chemisty. :)
5:05 AM
@rcollyer, well I do at least know what valence shell electrons are. I taught hs chemistry at one point.
@rcollyer The "garden variety" chemistry, anyway. You get into nuclear chemistry once you're worried about things like K-capture...
@AndyRoss I saw that on your linkedin profile.
@rcollyer stalker ;)
@AndyRoss Does everyone at WRI link to Stephen Wolfram?
@rcollyer making undergrads run away screaming is a badge of honour isn't it? I reckon the minimum effort way to do that is to go through a proof of the Hohenberg-Kohn theorems. Undergrads just love that stuff!
5:07 AM
@rcollyer, I don't rightly know. He added me so probably so.
@OleksandrR I'd be happy with tears (of joy or terror), but screaming works too...
@OleksandrR The HK theorems are easy. Even the undergrads can understand them. The Kohn-Sham stuff, maybe not.
@AndyRoss Just a thought.
Abusing my pinning privileges again.
@rcollyer what is pinning?
@AndyRoss it fixes the starred comments on the list for 2 weeks.
It can also happen naturally, if a comment gets to I think 5 stars.
@rcollyer I see.
5:10 AM
@AndyRoss It happens "naturally" when @rcollyer happens to like stuff =)
@rcollyer yeah, H-K theorems are OK at an undergrad level IMO. But I'm in a chemistry department, so often they balk at simple calculus. I managed to name-drop Tikhonov regularization in a talk I gave the other week.
@yoda I've done that only twice. I try not to abuse it.
@rcollyer lol.. just messing with you =) In any case, the bit about 5 stars getting auto pinned is not correct
Aha, so it was you who unpinned the "kitchen sink" logo proposal... :P
@OleksandrR The senior phys students are discussing Rayleigh-Ritz now, so they could grasp it easy.
@yoda how many?
5:12 AM
It seems the Mathematica world is full of physicists. The stats world is still woefully under-represented.
@JM me? I don't think so... I don't think I've unpinned anything
@OleksandrR FWIW: I learned Tikhonov on my own; that definitely isn't taught in a traditional chemistry curriculum, much less numerics...
@AndyRoss I'd like to learn more stats, personally. I just don't have the time.
@yoda No, rcollyer... :)
@rcollyer it never gets auto-pinned... it has to be explicitly done by a room owner or one with mod access
5:12 AM
@JM no, I didn't. I liked the proposal.
@yoda Ah.
@rcollyer isn't that what Mathematica is for? You don't have to know all of the gory details to get a basic feel for things and use them.
Then who stole from the cookie jar? I mean, unpinned it...
@JM no idea.
Ah well, doesn't matter now...
@AndyRoss The docs don't give enough detail, for my mind. I'd like to see more info on when it is appropriate to use something.
5:14 AM
@rcollyer examples?
@rcollyer as in, can you be more specific?
@AndyRoss The docs for QRDecomposition always seemed a little lite. But, I think it may have more details than some. I'd have to go back and look.
@Andy: I'd say that's why books like Wagon's are popular. Often the examples in the help file are minimal. Those books show off the power of these functions.
@rcollyer at least for my functions I put in as much as I can. I usually have to be reigned in :)
That's good news.
@AndyRoss I guess control theory and wavelets could benefit from some longer writeups. Those aren't areas I've got any familiarity with and without getting out a textbook or three it seems a bit difficult to find a way in even though they almost certainly do have uses in more general contexts.
5:17 AM
@rcollyer I'd definitely agree on those two.
@rcollyer its sort of a man power problem with those.
I wouldn't mind even external references at some points, also. A pointer to a book or even mathworld wouldn't be bad.
@AndyRoss always is.
What I'd really like to see is a short demonstration comparing, say, QRDecomposition[] and SingularValueDecomposition[] with LeastSquares[]. I could write one myself, but such things really should be in the help file...
@rcollyer most people are shocked at just how few people actually comprise the various teams that put this stuff together.
I agree. Even when it's documented, there are no references for the methods used. One thing I really like about MATLAB's documentation is that they meticulously include references to literature, so if you so desired, you can actually re-implement the algorithm yourself
@AndyRoss I worked in the software industry, so probably not.
5:20 AM
@yoda Yeah, I love how MATLAB points to articles.
Mathematica pointing to articles only happened very recently.
@JM Maybe that should be a goal for the blog.
this is another one of those man power issues. We are sort of limited in what we put in the home page for a function. For links and such we need advanced documentation which takes a lot of time.
Yes, that is definitely the case. Advanced level docs are time-consuming to create.
Yeah, you need to put a lot of thought into it, short of writing a damn textbook...
(I have a similar problem with writing math.SE answers...)
Yeah, you make others write them for you.
5:25 AM
@rcollyer :P
It is very late here, and I have to get up and write a presentation tomorrow. So, good night all. It has been good talking to you.
Most of the time is spent creating new functionality. I think most probably prefer it that way :) Then there is the time writing the docs which take a lot more time than I would have guessed. Advanced documentation keeps lingering on the todo list and never seems to rise to the top.
@AndyRoss I suppose, but would you really use a buzzsaw with only a barebones manual?
ok, JM was more colourful
@yoda *bows*
5:31 AM
This is true... but there are certain synergies in creating functions that are used to build more functionality...
And we developers are generally good at answering any questions that come our way :)
Well, you need a happy medium...
@JM we certainly shoot for it. Apparently we miss the target on occasion :)
among the WRI folks here (including undocumented folks), is there a way of knowing (or perhaps displaying in your profile) which areas you work on?
Kernel developer isn't terribly helpful >.<
@yoda we aren't supposed to give more than a job title... I stretched this as much as I was comfortable.
oh well.. alright. Was worth a try :)
5:41 AM
This makes me wonder how many "undocumented" WRI folks I'm currently speaking with :)
@AndyRoss Ah, so you can't say "I'm the special functions guy" or "yeah, I handle image processing"...
@JM I've said too much! ;)
@AndyRoss So now you know how we feel when we all get Internal`Blindsided[] with all the undocumented functions that get thrown around like candy on Halloween =)
@yoda don't think that only occurs outside the company. I find internal functions almost daily that I didn't know existed.
That's going to be an interesting blog feature: "Functions they hid"... :P
@AndyRoss Uh-oh...
5:48 AM
@JM well... I can think of several times when I was in the midst of some project and "invented" some new function only to find a better version already existed in some buried context.
@JM one you are probably familiar with is Internal`PartitionRagged. I reinvented that wheel many times.
@AndyRoss well, now you can post that as a question and @OleksandrR will pull one out of the hat for you :D
"Algorithm from division B was in fact the elegant solution sought by those kids in division A..."
@AndyRoss Ah, I have used that at least twice. Tricky...
@JM I use it all the time...
So on another topic... I actually liked the question 500 posted that was deleted. It had to do with creating labels with names, email addresses and icons (I don't know how to link to it).
@AndyRoss As you might have seen, I needed to generalize it in another question... :)
@JM I hadn't seen it yet I was busy here :) Obviously multitasking is not my forte.
@JM I see it. Thanks! I was puzzled by that one.
6:00 AM
@AndyRoss I suppose I should have mentioned the Internal` function, but I think it was more fun how I presented it... :)
@JM got me there, of which Internal` function do you speak?
@AndyRoss I mean, combine Internal`PartitionRagged[] with PadRight[]...
(Full disclosure: the code in that question was a modification of a "Mathematica Programming Competition" winner in the early '90s. Silly me forgot to note the year in the documentation.)
@JM I must not be looking at the same question then. Which question are you talking about?
@AndyRoss This one.
@JM Oh... gotcha. How does one link in the chat?
6:05 AM
So, [this](http://mathematica.stackexchange.com/) produces this.
@JM let me try it... I was talking about this
Ack, we were cross-talking... :D
I think he misunderstood the point of "ShrinkWrap"...
@JM apparently so.
Well all, I'm off to bed. It was good talking to you!
Likewise. See you Andy!
7:02 AM
@AndyRoss Is Internal`PartitionRagged new in version 8?
1 hour later…
8:02 AM
@JM I haven't used Excel since Office 2003 -- is that old or new? What is a better spreadsheet program?
8:47 AM
@MrWizard I really don't know. OO's spreadsheet is less than stellar; Excel is a steaming turd; and I haven't really needed to use spreadsheets in the last three years to bother shopping around. If you find a good one, please let me know, too...
(but 2003 is relatively old these days.)
1 hour later…
10:02 AM
If it's Thursday, it must be Verbeia-reports-on-traffic day
We are starting to get good engagement from newer users, but traffic is levelling out.
3 hours later…
12:45 PM
@Wizard I've been using Excel for the last 15 years (standard calculation tool where I work). From 2003 to 2007 there are substantial improvements. From 2007 to 2010, not so much. I recommend you see these two articles: add-ins.com/Excel%202003%20versus%202007.htm
Have little experience with the openoffice equivalent, but when I tried it (2 our 3 years ago), I felt it was a product of a rapid development project, trying to get it close to the functionalities of Excel as fast as possible, but where no time was taken to soften edges (for me, being obliged to use a spreadsheet all the time, it clearly justifies paying the small price of Excel)
@PFonseca thank you.
Also, if you intend to @notify me, make sure to use @MrW as @Wiz won't work.
Hello sjdh
Are the fonts very small on my plots, did some dpi setting get accidentally changed here? Can someone please see? i.stack.imgur.com/5DRoz.png
1:29 PM
@Szabolcs I use larger-than-default fonts (size 15) in GraphicsBox so I would not be the best one to ask. Also, regarding the filled curve answer I was slowed down because I don't actually have FilledCurve. It appears that my post-and-then-update approach is annoying however, and I will try not to do that as much. In this case I was going to leave out code and illustration entirely because I cannot generate it, but then I decided to copy from the documentation.
@MrWizard Oh you misunderstood me. A correct answer is always a good answer!
@MrWizard I do that on occasion as well.
But I'll try not to.
@MrWizard I never write @MrWizard, I only use @, M and Tab. Otherwise I would always have to lookup @Szabolcs name.
1:44 PM
@halirutan are you using a script for that?
All, I would appreciate feedback:
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...

1:55 PM
@Szabolcs you children show no respect these days...
00:00 - 14:0014:00 - 00:00

« first day (29 days earlier)      last day (3626 days later) »