« first day (278 days earlier)      last day (4196 days later) » 

12:25 AM
@JohnFultz Last one for tonight:
Can you explain what happens here?
g[x_] := Sin[x]
Definition /@ {g, g}
The output does look wrong in at first glance but then you notice that the output-cell is just really wide.
 
@halirutan The definition is being typeset as a single-column GridBox with the width of the column being set to Scaled[0.999]. In this context, Scaled is going to pick up its width from the window's width. Since there are two of them in this expression, it's always going to double the width. Some open questions here...
* Why isn't a line-break happening between the two GridBoxes? Hmm...looks like it could be a bug to me.
* Why is the column set to that width in the first place? I'm pretty sure there's a good reason for it, but I don't know what it is off the top of my head.
 
It's a really nice bug, because at first you think the Map goes completely wrong.
 
* How to work around it? This is merely a typesetting issue, so putting this in some other form would work. For example, InputForm.
 
Yes, it does does solve it.
Ok, I think I'll let you work and I promise I don't send FE-related bug reports tonight.
So you can have at least a bit Sunday.
 
Here's a more straightforward case:
Row[{"{",
Grid[{{1}}, ItemSize -> {Scaled[0.999], Automatic}, Frame -> True],
",", Grid[{{1}}, ItemSize -> {Scaled[0.999], Automatic},
Frame -> True], "}"}]
 
12:40 AM
yes, I look up the box expression of the output and there is the Scaled[..] width.
 
Or even...
Row[{Grid[{{1}}, ItemSize -> {Scaled[1.], Automatic}, Frame -> True],
"aaaaaaaaaaaaaa"}]
Replace 1. with 0.9 in this case and line-breaking happens as you might expect.
I'll go report it.
 
@JohnFultz Thanks. Then @István hasn't to report it, since he brought the attention to this.
Ok, I'm out for now.
Bye all and thanks @JohnFultz for everything.
 
I'm off, too.
 
1:32 AM
@chris Try this: Blend[{{0, Black}, {1/9, RGBColor[32/97, 1/62, 0]}, {2/5, RGBColor[44/59, 23/78, 1/32]}, {3/5, RGBColor[84/85, 1/2, 4/51]}, {2/3, RGBColor[84/85, 53/87, 1/10]}, {4/5, RGBColor[84/85, 4/5, 11/32]}, {9/10, RGBColor[84/85, 68/75, 46/75]}, {1, White}}, #] &
See if you can tell which one was produced with the original gradient, and which one was produced with the Blend[] function I gave above:
 
2:01 AM
lol: MathLink`CallFrontEnd[FrontEnd`UndocumentedCrashFrontEndPacket[]]
4
 
2:20 AM
@rm-rf It just had to be named that way... :)
 
2:41 AM
@rm-rf Haha
 
Well, I was secretly hoping for FrontEnd`UndocumentedMultipleUndo[] ;)
 
@rm-rf Hah
I'm actually quite happy with the SimulateMouseClick and family
thoguht I havne't tried to guess how they work, or if they work yet
I can remember at least a couple of times when I've wanted exactly that
 
@rm-rf I understand. Unfortunately, it is a very big job as it likely touches on everything in the front end. Massive amount of work, and ickiness.
 
 
1 hour later…
3:58 AM
@Rojo, you were asking about inverting SPD Toeplitz matrices before, yes? Have you seen Trench's paper?
 
@J.M. no I havne't
I'll take a look if I can access, thanks
 
 
2 hours later…
5:55 AM
@J.M. certainly more compact! and no I can't really tell. Thanks.
 
 
2 hours later…
7:43 AM
@halirutan , @JohnFultz: thanks for the analysis, and the report. I came around this bug by trying to put a function definition in a PlotLabel, like Plot[x, {x, 0, 1}, PlotLabel -> Definition@f], which leaves the label blank.
ToString and InputForm works, but not TraditionalForm on this one.
 
 
3 hours later…
acl
10:39 AM
all: can anybody run Plot[Sin[x], {x, -2, 2}][[1, 1, 3, 3, 1]] // Shallow and see if they get a list of points or an error?
(this is related to the discussion in the comments here
 
And btw:
Plot[Sin[x], {x, -2, 2}][[1, 1, 3]] // Length
gives 2
@acl What is in the 3rd part in your Graphics?
Ah, I see, I read the part in your answer. Hmm, seems strange.
@acl I get the error on Linux64 and on MacOSX64.
 
11:07 AM
@chris BTW, I'm still working on the Perlin thing; I'm not getting the artifacts you say you got.
@acl ...I still think it's safer to use Cases[Normal[Plot[(* stuff *)]], Line[l_] :> l, Infinity] // First to get the points.
 
acl
11:26 AM
@J.M. yes... I thought I'd avoid complicated-looking structures and this is the result!
I'm curious what is different on my system though
 
@J.M. that's using mma 8.0.4?
 
11:41 AM
@chris Yeah, the computer I'm on is has Mathematica 8.0.4 . I'm doing a line-by-line comparison now...
 
 
1 hour later…
1:07 PM
@halirutan, it seems that my estimate was very pessimistic... :)
 
1:20 PM
@J.M. 3 left ;-)
 
1:48 PM
@halirutan ...and, 10k. 'grats.
 
 
1 hour later…
3:13 PM
people really should be closing questions like this one... there are several such questions
Just because the original is "How do I convert A to a list" and someone asks "How do I convert B to a list" doesn't mean it shouldn't be closed as a dupe/TL when the answer is use Apply or use foo -> List
 
3:28 PM
@J.M. Thank you.
 
 
1 hour later…
4:30 PM
@rm-rf Believe it or not I cannot seem to find a good, simple duplicate for closing. You seem to do much better with searching than do I; do you have one in mind?
 
@rm-rf But in this particular case JM gets to bring in CoefficientArrays. So I like having it separate.
 
Hello all
@Mr.Wizard
 
@Rojo Hi.
 
Hey, how's all
 
4:45 PM
@Rojo, can you explain how exactly you do the traffic snooping of the FE?
The question is, how is it possible to listen to traffic of an already established MathLink connection?
 
Oh, no
I use LinkSnooper. It's just a different kernel
which, as soon as you open it
you get a java consolw eindow with all the packets coming and going listed
Quite practical. Once you set it up it's only a matter of choosing to use that kernel
Wait
 
Ah, I see it in the doc now.
@Rojo Got it up and running. Thx.
 
:)
 
5:14 PM
@Mr.Wizard a bit busy now and not at a comfortable computer, but I remember this mathematica.stackexchange.com/q/10833/5 was closed as TL because apply is basic. I'll getback on this later when I'm free
 
Anybody for a Calculus 101 problem?
17
Q: Limit of $\log (\log( ... \log((n)^ {(n-1)^ {....}})))$

belisariusThis is a spinoff of this question Defining $$f_0(x) = x$$ $$f_n(x) = \log(f_{(n-1)} (x)) \space (\forall n>0)$$ and $$a_0 = 1$$ $$a_{n+1} = (n+1)^{a_n} \space (\forall n>0)$$ How to calculate $$\lim_{n \to \infty } f_n(a_n) $$ (an "experiment" here, but (beware) I think WolframAlpha...

 
@belisarius I'm opening the link, with 0 hopes of being able to help
 
5:29 PM
@Rojo :(
 
@belisarius Yeah
I got no idea how to do it, but for some reason it doesn't seem too hard
if that makes any sense
 
@Rojo I thought like that too. But nobody came up with a single idea
@Rojo There was a site which gives you tentative formulas for a given number. Do you remember the name of it?
 
Nope
 
thnx!
 
No worries
That's what I'm here for
 
5:54 PM
@Rojo Are you still here?
 
Ya
 
:-)
I finally saw the problem with my "injector pattern" that you tried to point out months ago. I don't have a good grasp of the automatic scoping protection I guess. I think I have another method that is unaffected, but I'd like it, if you're willing, if you would lay out your understanding of the mechanism.
 
Ah
 
hello; I have a silly question: there is no way to make ListLogLinearPlot do ListLogLinearLinePlot is there?
 
Sure
 
5:56 PM
(sorry, slow typing; my IBM Model M finally broke and I hate these damn rubber dome types)
@chris Joined -> True?
 
@Mr.Wizard: arg ! It was indeed quite silly !
 
@Mr.Wizard
 
@chris it's the second example in the help. Are you making use of that?
 
the only mechanisms I am aware of that don't care about inner scopings
 
I am sorry. I used to use ListPlot[... ,PlotJoined-> True]
 
5:59 PM
are # functions and replacement rules on the lhs
 
this was made obsolete and I assumed ListLinePlot was the replacement
I missed the Joined-> True
 
@chris No need to be sorry. I'm just stating that if you don't read the documentation you'll miss all sorts of options.
 
All the others do care, and most of them (not all) behave the same way
which is
if you try to inject anything inside another scoping construct, the inner scoping construct variables get renamed
So, x /. x_ :> Hold@Module[{y}, x] gives Hold[Module[{y$}, x]]
but Hold@Module[{y}, x] /. x :> x doesn't give a crap
nor Hold@Module[{y}, #] &@x
 
I guess part of my confusion is that this works:
1 /. xx_ :> Hold@With[{xx = 1}, {1, 2, a}]
meaning that xx inside With is replaced.
 
Well, yeah
 
6:04 PM
@Mr.Wizard Hi Mr!
 
@belisarius hello
 
@Mr.Wizard @Rojo There was a site which gives you tentative formulas for a given number. Do you remember the name of it?
 
You don't mean OEIS do you?
 
Sorry, @Rojo, I was attempting to quote
@Mr.Wizard nooe
nope
 
Its ok
 
6:05 PM
Given ONE number
 
Ah.
 
one REAL number
 
Wizard, how they behave when you actually want to replace the inner variable, that's some of those things that depend onf what you are using to inject
 
something like reverse lookup formula
 
patters do omne thing, Module does another, etc
but all of them I think, rename the variable when youa re trying to inject anything inside the scoping construct (not changing the variable itself)
 
6:07 PM
@belisarius I'll let you know if I find something like that. Sounds useful.
 
@Mr.Wizard yep, the other day someone mentioned it here
 
I know that site
I just don't remember it
 
Thanks Rojo. I should know this as I recall, because Leonid wrote about it before, but my memory is not great.
 
I also know that wolframalpha does it "sometimes"
 
@Rojo some keyword?
 
6:09 PM
No, nothing comes to my memory... You could google the keyword WA uses for that "possible closed forms". Who knows
 
@Mr.Wizard it's ok, we all forget. And I hope I said the same as Leonid, because I don't remember having learned it from soemthing I read
I remmeber having been confused by forum posts, and playing around to get the idea
 
@Rojo It worked
inverse symbolic calculator
@Mr.Wizard bookmark it
 
6:15 PM
Actually, is that the same one?
 
@Mr.Wizard yours is better
but hangs with "my number"
 
Figures. :^)
 
7:04 PM
@belisarius where did that question come from?
 
@Rojo from here
10
Q: Preventing Numerical value from being evaluated

CBenniI have searched stackoverflow (and comparable pages) for quite a while now (got redirected from there to this specialized stack), and I surrender. I am trying to evaluate an expression that is small in the end numerically. Example: Log[Log[Log[6^5^4^3^2^1]]]=12.9525... WolframAlpha has no pro...

 
 
2 hours later…
9:19 PM
Anyone familiar with the GitHub StackMMA/Attachments stuff? @rm-rf? @halirutan?
 
@SjoerdC.deVries aehm, yes, I should be. Why
 
I want to dump a file there but couldn't figure out how
 
@SjoerdC.deVries, give me a sec.
 
help seems to be down, and I only see buttons related to downloading not uploading stuff
 
@SjoerdC.deVries I always do this from console like in cvs or svn.
I have the directory tree on my system and just commit and push files to the server.
 
9:25 PM
@halirutan Unfortunately, I don't have that
If that's the only way, I'll fall back to my code-in-picture trick.
 
@SjoerdC.deVries and http://code.google.com/p/tortoisegit/ is no option?
I never used or even know that there is a webinterface for a git repo where you can upload stuff.
 
@halirutan I've never used it, but someone (I presume rm -rf) gave me access to it so that I could move attachments to a more secure place. I assumed this was an easy option, but if have to install all kinds of programs and have to familiarize myself with all kinds of complex procedures this isn't really an attractive option.
 
@SjoerdC.deVries It's still a versioning system. It assumes that either someone wants to only download all or parts of the repository for compilation and if someone wants to upload files, he is developer who should have all/parts of the source-tree on his computer.
Are you on Windows or on Linux or on Mac?
 
@halirutan Windows
I just want to upload a simple .m file
 
Then the Tortoise-tool in the link are really simple. They attach to the explorer and you can throw stuff in and upload it by click.
(At least thats how they work for cvs)
@SjoerdC.deVries Wanna send me the file? I upload it?
 
9:45 PM
@SjoerdC.deVries I don't upload by committing because I'm not sure how to manage it with so many users (each will have to pull/fetch regularly and keep files even if they don't need it). If you go to the downloads page, there's a link to upload
 
@rm-rf Where?
Ahh.. under Downloads
very clever
 
@rm-rf @halirutan I already used the picture trick, a process that takes less than a second. I'm off to bed now, but feel free to remove the piece of code and replace it with a link to GitHub.
0
A: What do AstronomicalData's "SetTime" and "RiseTime" values refer to?

Sjoerd C. de VriesA couple of years ago I was in an email conversation about this topic with Jeff Bryant, a WRI employee. He was not directly responsible for AstronomicalData, but he told me that Mathematica did not correct for atmospheric refraction. Good to know, as the refraction at the horizon is about the sam...

 
@halirutan yep:
@SjoerdC.deVries You should see something like the above on the Downloads page... you can easily choose a file and upload there
 
10:02 PM
btw, congrats @halirutan :)
 
@rm-rf thanks.
 

« first day (278 days earlier)      last day (4196 days later) »