Wolfram Mathematica

1:07 AM

@halirutan @RunnyKine Are you actually doing anything interesting with CUDALink ?

1:39 AM

@RolfMertig, I'm trying to see how feasible it will be to incorporate it in some molecular simulation analysis stuff. Haven't used it before though.

Rolf Mertig

1:58 AM

@halirutan sorry, pinged you by accident about CUDALink ... Gruss aus Mexico

halirutan

@RolfMertig Was about to answer you.

@RolfMertig But no, I'm not doing anything with CUDALink. It never worked out of the box in Linux although I always have set-up everything for CUDA-dev. Since it was introduced you always had to manually fix some library paths etc which is just annoying. The next point is, that I couldn't simply share the code since most of my colleagues haven't installed it.

(by haven't installed it I mean, they have no usable graphics card)

(just got up, so I'm not at 100% perf)

Rolf Mertig

oder Kaffee!

(to boot up)

halirutan

@RolfMertig Oh, CUDAQ just crashed my V10.

(and of course returned False)

Rolf Mertig

CUDAResourcesInstall[Update -> True]

halirutan

@RolfMertig Running. Progress -> 0.000014%

(taking a shower. Back in a few to check)

Eric

2:12 AM

hi

Rolf Mertig

hi

Eric

Mathematica gave the correct symbolic solution of this double integral:
`Integrate[
4 - x^2 - 2 y^2, {x, y} \[Element] Disk[{0, 0}, {2, Sqrt[2]}]]`
Did Mathematica use Fubini's Theorem insde?

(this can just be done in version 10, not in earlier version)

Meta

3:32 AM

0

Q: CDF On iPad: vaporware?

Haven't sen any subsequent announcements after the 2012 hoopla about CDF on mobile platforms. Is CDF essentially a zombie technology? wolfram blog

Szabolcs

4:02 AM

@Eric Here's how to do it in v9: Integrate[(4 - x^2 - 2 y^2) Boole[x^2/4 + y^2/2 < 1], {x, -2, 2}, {y, -2, 2}]

Eric

Do MMA use the Fubini's as part of its algorithm?

In Caculus, there seems to have not other way to compute mutiple integral on strange regions

Since mathematica can evaluate the integral on ellipse, disk, or arbitrary region generated by ImplicitRegion,...

Yves Klett

11:38 AM

Oh dear, weather homework: Need sm help, plz!

uh, and already gone. That was quick.

Yves Klett

12:32 PM

@RolfMertig Mexico, eh?

acl

12:46 PM

@YvesKlett this also does not seem like a reasonable question...

Yves Klett

@acl true, dat.

Michael E2

@YvesKlett Been asked a couple of days ago

Yves Klett

@acl ...and closed as well ;-)

acl

@MichaelE2 more than once.

@YvesKlett Their previous question was similar and got an answer. This apparently was so over the top that it got closed.

Yves Klett

@MichaelE2 right :(

Michael E2

12:49 PM

I guess school is starting somewhere in London.

Yves Klett

@MichaelE2 U suR? Nds hlp plz!

acl

@MichaelE2 In August? I doubt it

Michael E2

@acl Yeah, seems early. Perhaps somewhere else. My daughter's school started at the end of July, but this place is nuts.

Yves Klett

@MichaelE2 the overall state of grammar would be very depressing for London, indeed.

acl

It is probably people who, in their innocent thirst for knowledge, turn to mma.se as a fount of wisdom

@MichaelE2 end of july? when did they close?

Yves Klett

12:56 PM

@acl you are a philanthropist

Michael E2

@acl End of May. They used to call it the "year-round" schedule, but it doesn't go year-round. There's a week off every six weeks or so.

acl

@MichaelE2 oh OK, 2 months then, that's not bad

Michael E2

@acl Yeah, but they tend to start on the hottest day of the year. (Georgia, USA)

acl

@MichaelE2 Well, knowledge can only be gained by hard work and sweat

alternatively, the moment a random question occurs to you (eg, "please determine a guitar's eigenmodes for me"), ask it here

someone's going to answer!

Michael E2

@acl It would make a good Wolfram Blog post, though. If it's computable, that is.

acl

1:10 PM

@MichaelE2 I don't know about this question but the last (the eigenmodes of the wave equation on a simpler-shaped 2d membrane) took me a few days when I solved it a couple of weeks ago. This looks like it requires much more work (either treat the thing as a 3d object, in which case I don't know if it's practical to solve in terms of computing power using this method) or somehow incorporate the extra structures in the (already complicated) 2d shape.

Either way, I find it completely unreasonable to ask this directly

Michael E2

@acl It seems that way to me, too, but user21 seemed interested in the question.

Yves Klett

@MichaelE2 well, user21 does like finite-element-based questions :D

Michael E2

@YvesKlett And I like reading his answers. :D

acl

@MichaelE2 I'm going to take a wild guess that he has written a PhD thesis on this sort of thing, though!

Yves Klett

@MichaelE2 absolutely.

@acl and yes ;-)

m_goldberg

1:23 PM

Do you think when school starts some student sends all his/her friends a text message saying "Hey, I found this neat site where some geeks will answer all our Mathematica homework questions"? Well, they wouldn't write it out like I did, but I don't how to code in texting-speak, and I'm too old to want to learn.

Michael E2

@m_goldberg Yes. I think it's quite natural.

m_goldberg

1:38 PM

@MichaelE2. Like a lot of others who participate, I'm willing to help students with their homework if they formulate the questions reasonably. Unfortunately, many seem to be clueless about how to ask a technical question,

Anybody. Do you think it's part of our mandate to teach question posters how to ask good questions? Should ask this on meta?

Yves Klett

@m_goldberg this must already have been discussed somewhere on SE?

m_goldberg

@YvesKlett. Yeah, I suppose you're right. I search on it.

@YvesKlett. I found this. Very interesting. I didn't know the "homework" tag had been officially deprecated.

rm -rf

2:05 PM

18

Q: Policy on Homework questions

As we have received two homework questions today, and likely will have more in the future, what should our homework policy be? Should we ban them entirely? If not, what degree of help should we give? Also, do we make a distinction between Mathematica being used to solve a homework problem, or the...

discussion tagging

m_goldberg

@rm-rf. Excellent. Am I correct to presume that this site can have a homework policy independent of SE as a whole?

xslittlegrass

Hi, I have a quick question and feel like it should be answered before but I can find it, could someone help? Why ListPlot or ListDensityPlot do not speedup even if plot only a very small region? For example, compare

rv = RandomVariate[ExponentialDistribution[2], 2*10^5];
t = AbsoluteTime[]; ListLinePlot[rv, PlotRange -> {{0, 10}, All}]
AbsoluteTime[] - t

and

t = AbsoluteTime[]; ListLinePlot[rv, PlotRange -> Full]
AbsoluteTime[] - t

I get 1.41 second for the first and 1.37 second for the second.

Yves Klett

@xslittlegrass PlotRange only influences the viewport - all points are still there, just not visible.

xslittlegrass

2:22 PM

@YvesKlett you mean they all get rendered but not shown?

Yves Klett

@xslittlegrass no clue about the internals, but all are processed somehow. Probably the rendering itself does not take a lot of time for multi-line-primitives as used by ListPlot etc.

It seems the clipping by PlotRange takes more time than the rendering of all elements in this case.

xslittlegrass

@YvesKlett OK, I'm not sure I understand correctly, but from this post "ListPlot: Plotting large data fast", rendering seems to take the majority time.

for large data set

@YvesKlett do you know why it has to go through all the process? Or what's the advantage of doing that?

Yves Klett

@xslittlegrass perhaps 10^5 points are not large yet...

@xslittlegrass it may be more effective to do the sampling/selection beforehand and then plot.

xslittlegrass

2:38 PM

@YvesKlett OK, thanks for the help :)

PlaysDice

Can anyone guide me if M has the special symbol 'eet' (basically an upside down T)? I cant see it in the documentation

hah! (random typing later).... its [esc] perp [esc]

Szabolcs

3:00 PM

@Eric I don't know how Integrate works.

Karsten 7.

@PlaysDice There is also ⊥ [UpTee] (Esc uT Esc)

PlaysDice

@Karsten. Thanks. uT is smaller, for sure. As far as I can tell, perp isnt listed as a named character. +1 for the random monkeys!

RunnyKine

I'm getting an error when trying to save a really large (~2GB) Dataset to a binary (.mx or .wdx) file. It appears to have saved correctly but when trying to read it back, I get an error saying the binary file is corrupted. Has anyone noticed such behavior?

rm -rf

@m_goldberg yes, certainly. While we use meta.SE as a general guidance, we're free to make our own rules (within reason) if it is better for our community.

Pickett

3:18 PM

I've noticed more and more downvotes lately. This post lacks effort perhaps but three downvotes and no explanation? That's not good.

Yves Klett

@Pickett well, I hinted at missing code and the inconveniece of that.

RunnyKine

@Pickett, I agree. I got a random downvote on one of my answers the other day but whatever.

acl

@Pickett this has 3 upvotes, though, so it's all good!

Yves Klett

@RunnyKine just had a look - strange that. But most likely just a coordination mistake :D

acl

@YvesKlett It could be, I once downvoted @Szabolcs by mistake (on an old iPhone)!

Pickett

3:24 PM

@YvesKlett That helped him some, but people who are new to the site might not connect a polite suggestion and the downvotes which in my mind are more aggressive as in "you shouldn't have posted this"

RunnyKine

@acl That's funny.

acl

@RunnyKine I only found out because he came here to whine, I went to the answer out of curiosity and noticed that it was I that had downvoted him...

Yves Klett

@Pickett today saw a slew of bad questions... I am always happy to reverse a vote if the question is edited.

RunnyKine

@YvesKlett, you're probably right.

Pickett

@acl I think there are people who upvote questions they would like to see answers to, even if they aren't good for the site. I personally don't try to counteract such votes, but maybe that question is so out of bounds it deserves downvotes in itself. I can certainly see a case for downvoting.

acl

3:26 PM

@Pickett It does seem a bit too much. But maybe I am too grumpy

Pickett

@YvesKlett Alright, fair enough

RunnyKine

@acl, so did you correct the mistake? I know votes get locked in after a while.

acl

@RunnyKine Yep

Pickett

Anyway I'm still learning how to use the voting system and felt like making a general remark. I wasn't looking to criticize anyone, I'm sorry if it came across that way.

Yves Klett

@Pickett no worries - I am constantly reconsidering my downvote/closure motivations, and they are not always consistent, and quite often really subjective.

acl

3:52 PM

@Pickett did not look like criticism

and anyway discussion is good

Szabolcs

4:19 PM

@TaliesinBeynon Regarding your question about what features we'd like to see in Mathematica: making it reliable and fixing bugs should really really be the top priority. It's possible that I push the system harder than most people, but recently no week passes by without finding some problem, and it can really be hindrance and a time waster ...

Szabolcs

5:01 PM

A close second to that would be performance, which is often really lacking. Any computation with units is extremely slow, and now all the built-in datasources, as well as some new functions use units. I know lots of people have been asking for units, but how many people do really use them for serious work? I'd guess very few. Probably the only reasonable application is within the classroom.

All because of the awful performance.

kirma

Is there a name for this or reasonably similar problem? There are N sets of natural numbers. Find an alignment of these sets so that when every set is assigned an natural number "offset", summed to its' values, all numbers over all sets are unique, and difference between minimum and maximum used offset is minimized.

xslittlegrass

Hi, I have a quick question. How can I get (! FreeQ[#, 1]) && (! FreeQ[#, 2]) & from a list {1,2} ? Thanks !

Szabolcs

@xslittlegrass What is #? A flat list or a nested expression?

Not@FreeQ[#, 1|2]&, i.e. Not@FreeQ[#, Alternatives @@ {1,2}]&.

xslittlegrass

5:16 PM

@Szabolcs I want to select some expression that contains both 1 or 2. For example I have a list {-c, a d, b d}, and I want to select out terms that contains all terms of a list {a,d}, so that it selects out a d.

@Szabolcs Select[{-c, a d, b d}, Not@FreeQ[#, Alternatives @@ {a, d}] &] gives {a d, b d} instead of {a d}

Szabolcs

sorry about that

@xslittlegrass so your expressions are not flat lists then?

xslittlegrass

@Szabolcs yes

Szabolcs

terms = {a, d};
list = {-c, a d, b d};
Select[list, Intersection[Level[#, -1], terms] === Union[terms] &]

@xslittlegrass ^

Works only if a and d are symbols ...

xslittlegrass

@Szabolcs Thanks. What if I want to select terms that contains Derivative? For example

terms = {Derivative, d};
list = {-c, d'[t], b d};

Szabolcs

@xslittlegrass Sorry, I can't build anything better right now than your original (! FreeQ[#, 1]) && (! FreeQ[#, 2]) & expression, or equivalents of it.

xslittlegrass

5:29 PM

@Szabolcs OK, that's fine, thanks for the help :)

brama

@halirutan or @Szabolcs or anyone else: Any one with experience running MMA on Linux Clusters using bash/PBS file?

xslittlegrass

@Szabolcs I found one solution !

terms = {Derivative, d};
list = {-c, d'[t], b d};
f = Function[{t}, And @@ ((! FreeQ[t, #]) & /@ terms)];
Select[list, f]

rm -rf

5:51 PM

15

A: Speed of curated data calls in Version 10

There are system options available that should restore the old behavior for most of the currated data paclet: SetSystemOptions[SystemOptions["DataOptions"] /. True -> False] {"DataOptions" -> {"ReturnEntities" -> False, "ReturnQuantities" -> False, "UseDataWrappers" -> False}} Note th...

Might solve some of the issues with data sources using units, but old functions returning units and being backwards incompatible is still a HUGE pain.

RunnyKine

6:07 PM

@TaliesinBeynon sorry to bother you with this but I'm getting an error when trying to save a really large (~2GB when saved as text ~9GB in memory) Dataset as a binary (.mx or .wdx) file. It appears to have saved correctly but when trying to read it back, I get an error saying the binary file is corrupted. I have now tried this on three different computers with the same outcome. Is this issue known?

Vitaliy Kaurov

6:39 PM

http://community.wolfram.com/groups/-/m/t/313447
New book in CDF format: Interactive Computational Geometry

Vitaliy Kaurov

10:48 PM

Community questions with no replies:

http://goo.gl/McJoC5

Feel free to forward to StackExchange if they were answered here.

Thanks!

Mr.Wizard

11:24 PM

Anyone who cares about site mechanics and hasn't yet voted on this issue, please do:

15

Q: How should we tag longstanding bugs that have been fixed?

Using What to do with [bugs] questions now that version 9 is released? as a guideline Questions about bugs that have been fixed should have individual tags for each affected version. However this is impossible for longstanding bugs where there are many affected versions as a Question may only ha...

discussion featured tagging

I'll be Accepting and implementing Szabolcs's solution soon unless someone poses a counter-argument or it suddenly starts receiving down-votes.

halirutan

@Mr.Wizard I upvoted Szabolcs solution and I'm for using a fix template to be able to search like Silvia suggested. I guess if I would search for an issue, I would just search for [bugs] and the topic and then read through whether it is the issue I'm having. With this I instantly see whether or not it has been fixed.

Pickett

Just today a new situation arose here - a bug that was fixed but not in any update of MMA :)

I support Szabolcs' solution as well. Let's put it into practice...

Honestly I don't know if the post I linked to should have the [bug] tag because since it was fixed on the server side it cannot happen to anyone. It's just a historic factoid that "for the first couple of months Permissions -> Public didn't work properly for some functions"

Transcript for