Wolfram Mathematica

8:07 AM

@MrWizard Are you sure it was a good idea to close this? Code review and code golf are two very different things. Your original question only received answers about code golf, and when the new question was closed, it effectively shut down discussion on the code review issue.

Of course I'm not asking for it to be unmerged. That would not make any sense at this point. But I think in particular instance closing it didn't have any positive effect.

Mr.Wizard

8:36 AM

@Szabolcs I am sorry that it had a bad effect, but I still do believe it was a duplicate. I really am not sure how it could have been handled better. I guess I could have edited my question to only apply to "code golf" but that was never my intention when asking. I see that you never edited my question to address your specific concerns. Would you please try that?

Szabolcs

Of course that's what I'll try if/when the issue comes up next time.

@MrWizard Isn't it half past four there? :-)

Mr.Wizard

1:40AM Pacific Daylight Time

I really am sorry that my merge ended the discussion, if that is indeed the case. :-/

Szabolcs

Oh, I was confused about your location then

Mr.Wizard

here :-)

Szabolcs

It's not a big deal, of course. I just had the impression that de facto it wasn't treated as a duplicate (i.e. in how it was answered), even if technically it was clearly a duplicate.

Oh, of course! I am confusing state names starting with O- ...

There was a code-review-like question yesterday

or today, your time

Looking at that meta thread, my impression is that people are against code review type questions

but looking at the question, I think it's a fine question

So maybe let's not have a tag code-review (which might encourage not-so-good questions), but keep good questions like the one I linked to.

Mr.Wizard

8:47 AM

I think you should edit my question to place emphasis on Code Review. How about announcing that question update here and pinning it?

Szabolcs

@MrWizard Maybe later, I don't want to promise I'll do that if I won't have time... :)

Got to go now (and you got to get some sleep!) Good night!

Mr.Wizard

Okay; you could cut&paste from your own question.

Good night, Szabolcs! I'll get my sleep sooner or later. :^)

Rolf Mertig

9:05 AM

@OleksandrR It is slightly involved, but looks like that for the problem at hand that very algorithm seems to be needed.

Szabolcs

9:51 AM

@MrWizard Nope, not asleep. It's 12 o'clock here :-) I need to go out, but decided against going now because of the heat ... (I'm in Malta)

Mr.Wizard

@Szabolcs right, I wasn't thinking; honestly time of day has little meaning to me.

Szabolcs

We're officially doing better than MathGroup

The mean monthly post count on MathGroup is ~840 for the past three years, based on this data: https://groups.google.com/group/comp.soft-sys.math.mathematica/about?hl=en-GB . If our A51 stats can be trusted, we have 12.3 post per day with an answer ratio of 2.3, which means `12.3 * (1+2.3) * 30 = 1218` posts per month. Also, because of the possibility cleanup, comments and edits, and the lack of endless duplicate answers, our post quality is higher than MathGroup's.

Mr.Wizard

I haven't been following MathGroup recently, but I expect it still is frequented by knowledgeable users with interesting posts. I have hoped that SE would free MG of some Q&A traffic and allow it to focus on more "philosophical" material.

Szabolcs

Another noticeable effect is that the MathGroup post frequency has decreased noticeably. It has not been as low as in last month since 2001 December. It shows that people are flocking over, but whether the decrease post frequency over there is good or bad.... it's not so clear.

Mr.Wizard

How is the post quality on MG these days?

Szabolcs

9:58 AM

I don't follow regularly. To be honest, mostly I've been showing off this site there by posting links from Mathematica.SE that answer questions perfectly.

My impression is that it's the same as it used to be. The main problems are lots and lots of duplicate answers because of the moderation delay and a few low quality beginner posts because the format is not really suitable for preventing those.

Occasionally there are answers that are better than ours, for similar questions.

Actually, I think I should make a note whenever I see that.

The enthusiasm is much greater here, and questions almost always get answers (which is not true for MG). Just the fact that it's usually easier to follow the conversation here makes for better answers in the end. But I wouldn't mind if a few more experts came over.

Mr.Wizard

It would be nice if MG turned into a kind of community blog for interesting topics, and the discussion thereof that is awkward here.

I know I'd be more encouraged to regularly read it again if that were the case. There was real gold there hidden in the daily noise.

Szabolcs

MG still allows some topics that just aren't suitable for SE. For example, announcing packages, or simple discussion about topics without the QA framework. I hope Mma.SE will not kill MG.

@MrWizard About the ticks question: the main question is how to put the tick labels on the right hand side of an axis instead of the left hand side, so they won't intersect with the plot. I'm not sure how to do this conveniently, but I think I've seen something like this somewhere ...

Take e.g. Plot[Sin[x], {x, -6, 6}], and note how the curve intersects the -0.5 label. He wants to move that label to the right, only below the x axis.

Mr.Wizard

@Szabolcs you mean like this? mathematica.stackexchange.com/a/2608/121

Szabolcs

@MrWizard Yes, you could link to that in a comment

Mr.Wizard

10:14 AM

Done.

He's got a graphic up now too.

Szabolcs

It's a very reasonable question ... your supervisor asks you to move the ticks to the other side and won't understand why you'd have any trouble with such a simple operation. For lack of time you might do it in Illustrator and hope you won't need to regenerate the figure ...

Mr.Wizard

10:51 AM

@Szabolcs is there a way to make Rotate do a simple translation with its extra arguments, or make Translate (or similar) operate on an arbitrary expression? I'm trying to some the tick label problem in this way.

rcollyer

1:35 PM

Morning!

Szabolcs

Morning.

rcollyer

Although, it is likely very late where you are. :)

Szabolcs

three thirty

rcollyer

9:30 here.

Szabolcs

or, as the locals say it, "half three"

which bothers me to no end because in Hungarian the literal translation of "half three" means 2:30.

rcollyer

1:38 PM

that's interesting

Szabolcs

"half past three" is okay, but "half three" means 2:30 to me. It's like that in Hungarian, in Norwegian and I'm sure it's the same in several German speaking areas.

rcollyer

we use "half past" here

Szabolcs

Here they speak a mixture of Arabic and Italian and they constantly code switch with English. No wonder some things get messed up in the language.

rcollyer

a friend from India would do that when talking with his father, except they'd use about 3 different languages, including english.

Szabolcs

You'll often here "ciao aleikum" on the street here. It sounds hilarious to me :-)

rcollyer

1:44 PM

That's funny, and a bit weird. :P

Quick digression, did you see yesterday's Astronomy Picture of the Day?

Szabolcs

(Just in case you didn't take the "mixture of Arabic and Italian" literally :)

rcollyer

apod.nasa.gov/apod/ap120603.html

Szabolcs

Nope

Checking

Oh, I missed that!

I totally forgot :(

rcollyer

I set it as my home page. :)

Szabolcs

June 5-6, so there's still time!

rcollyer

1:48 PM

Yep. The photographer for that one, though, had a very good eye. Brilliantly composed.

Szabolcs

2:07 PM

I contacted the local astronomy society, but as it's around sunrise, there are no buses yet, and of the assumption is that "everyone has a car". I really don't like the car culture here.

rcollyer

@Szabolcs Being a US native, I'm used to the car culture despite not liking it. I much preferred the time I spent in London or Atlanta where you could get everywhere you wanted on decent (and inexpensive) public transport.

Szabolcs

Here the whole island is barely 15 by 30 km, and the country has the 5th highest cars per population ratio worldwide. Also, imagine a tiny island with noticeable air pollution...

It'd be an excellent place to bike, but with this car density it's dangerous.

rcollyer

Reminds me of Los Angeles.

Noticeable air pollution, not helped by geography.

And, lots and lots of cars.

Heike

2:45 PM

@rcollyer The average size of american cars probably makes it worse.

rcollyer

@Heike It certainly doesn't help. Add to that, there is a strong lobbying group that stymied improving gas mileage and emission standards for years. On the other hand, California has the most stringent emissions standards in the US, and if a car will pass there, it will pass any where in the states. This is, of course, due to the problem in Los Angeles. :)

Heike

@rcollyer People are are actually lobbying against improved gas mileage?

rcollyer

@Heike Yes, and are well payed for it.

J. M.

Clearly people quite like their Humvees. :P

Heike

They're probably paid by the oil companies

rcollyer

2:53 PM

There is an ongoing debate here about the role of corporations in government, and is the corps that are paying for it.

@JM not just humvees, but SUVs, in general. Or, as a friend liked to call them: Silly Useless Vehicles.

J. M.

@rcollyer The problematic part is the treatment of corporations as actual people, with rights and everything... yech.

rcollyer

@JM In a nutshell, yes.

I am firmly on the side of "when a corp can be convicted of murder, then I'll consider them people" until that point ...

Heike

@rcollyer In 99.9% of the cases, they are a complete waste of space imho.

J. M.

@rcollyer Hell, if you're living in a place with less than perfect terrain, sure. But most of the buyers of these things travel on paved ground, and will likely throw a fit if even a speck of dust gets on their ride. Ugh.

rcollyer

@Heike I can see it with our toddler. His "equipment" requires a lot of space, and our car is a little undersized for it. Also, there are newer models of some that get better gas mileage than our car. :P

@JM That is the primary argument. My uncle-in-law likes to point out that if you need 4-wheel drive, you shouldn't be driving.

J. M.

2:58 PM

@rcollyer I thought minivans are the craze these days? :)

rcollyer

@JM most people don't like mini-vans, and my wife despises station wagons, so SUV it is (when we get the cash to upgrade, that is). We rented one for a long weekend a couple of years ago, and liked the model. That and the very good gas mileage, and there is a hybrid version out there (30 hwy, 33 city).

That's miles/gallon for the SI heathens out there. :P

Heike

@rcollyer I have a friend who owns a tiny farm in the peak district in England. A normal car would probably break down within a month just from getting to and from his house.

rcollyer

@Heike my parents live in the same sort of environment. lots of hills, plus equipment to haul. (my dad is a cabinet maker.)

J. M.

(I'm still laughing at that "save gas, save the environment!" bumper sticker attached to a Humvee...)

rcollyer

@JM that's funny. For there size, they don't get bad gas mileage, for their size ...

I should get something done today. Talk to you all later.

R.M

3:11 PM

throw this to the sharks at SO?

J. M.

I'm torn...

Heike

@RM I don't think it would be well received over there.

J. M.

(Good luck reimplementing FindClusters[] in VBA, tho.)

Heike

The first 6 should be easy (and if they're not then VBA sucks).

R.M

@Heike I don't either... but translation to VB will probably just stagnate here (and I consider it off-topic)

Heike

3:35 PM

@RM I agree that the question is off topic here, but I think it should be rewritten a bit first to give it a chance on SO.

Szabolcs

3:46 PM

@RM Yes.

@RM I voted to close as off-topic...

Oleksandr R.

@JM I don't think I ever really understood this argument, at least as far as it goes in the US. On the one hand it could be interpreted as a view that the rights of legal persons should be distinct from and a subset of those of natural persons, but I have never heard it couched in those terms. On the other, it could be interpreted more literally as a call for legal personhood to be abolished altogether, but it seems to me that only

the anarcho-capitalists have truly understood what this would actually imply and support it as such. I really can't see the Michael Moore crowd advocating what is essentially an anarcho-capitalist utopia. So, have I missed something fundamental?

@rcollyer that's already possible in some jurisdictions: see e.g. en.wikipedia.org/wiki/Corporate_manslaughter. Not certain why the US hasn't adopted similar legislation so far?

Szabolcs

Why do so many new users turn up on SO? --> stackoverflow.com/questions/10884034/… I often flag for migration, but it might depend on luck (and which moderator sees the flag) if the question ever gets migrated ...

J. M.

@OleksandrR My reading of it is your "on the one hand". The problem is that corporations have capabilities persons don't have, so the playing field isn't terribly fair if you give corporations the rights of persons as well.

@OleksandrR One word: lobbyists.

Oleksandr R.

@JM Corporations are legal people though, unless you want to discard nearly 900 years of English common law and start from scratch. For instance, I can't imagine that anyone would seriously argue that legal people should not be entitled to protection under the law and access to the justice system. However, certain rights of natural persons, such as voting, can sensibly be withheld from legal persons.

J. M.

@OleksandrR Yes, the withholding of certain rights is what they're rabbling about there these days, last I checked.

Szabolcs

4:01 PM

Anyone else wants to give this guy a push to register? mathematica.stackexchange.com/users/1405/anastasiia

Oleksandr R.

@JM I see, okay. I've never heard it expressed in that way, but at least the argument makes sense when put in those terms. It had struck me that abolishing the concept of legal persons and corporations altogether would essentially abrogate the modern concept of the state and all state apparatus such as criminal law, which perhaps is not what is really intended with that argument.

rcollyer

4:18 PM

@OleksandrR I am of the opinion that legal persons have no inherent rights, they have privileges that they must continually pay for (i.e. taxes), but they should be easily revokable. In a recent Supreme Court case, Citezen's United, corporations were given the right (with only a few caveats) to dump large amount of cash into our political system on the pretext that as they are just groups of people, then they should have the same rights as the individual.

J. M.

@rcollyer The rub is the "easily revokable" part; they pay big money to ensure these rights never get revoked. Ugh.

rcollyer

@JM except, the current legal doctrine is that they're rights, not privileges, so they are not revokable, and yes they pay very big money to ensure that it remains that way.

J. M.

@rcollyer Hmm, I missed that distinction in my reading on this stuff. In which case, even more ugh.

rcollyer

Reminds me of the cyberpunk literature genre: welcome to the future.

Oleksandr R.

@rcollyer interesting. Do you consider in this conception that the state is still a legal person and has the ability to press criminal charges, or does all criminal law become civil law?

rcollyer

4:31 PM

@OleksandrR that is an interesting question, and it has not occurred to me. Truthfully, I view government bodies to be distinct from corporate bodies, although, in some sense, the law treats them as the same. However, their purposes differ drastically, i.e. a corporation ultimately is intended for profit.

Oleksandr R.

Is, say, the president, immune from personal legal charges resulting from the actions of the government? Or must he/she defend themselves in court if any charges are brought? Who operates the justice system in this framework?

J. M.

I would think that rules for states and rules for corporations ought to be different...

rcollyer

@OleksandrR Currently, that is true, well, mostly true. The president can be charged with a crime, then Congress is empowered to try them for the purposes of removing them from office. To some extent, they're even shielded from personal law suits during their tenure.

Oleksandr R.

@JM the problem is, the state is a corporation, unless you have some figurehead through which all the powers of the state are exercised (e.g. king/queen/president). In the US cases are not brought as "President vs. X"; they are "United States vs. X" or "Name of State vs. X"

J. M.

"Congress is empowered to try them for the purposes of removing them from office." - i.e. impeachment.

Oleksandr R.

4:36 PM

Yes, but only Congress is empowered to do that. Ordinary citizens can't sue the president in civil court.

J. M.

@OleksandrR Huh, I thought it was "the people of (state) versus X"?

rcollyer

@JM I think that is technically true, but it is often shortened to "state v. X".

Oleksandr R.

@JM well, it doesn't much matter. "The people of X" is a legal person. If this is not valid then a specific person has to bring the case, not "the people" in general.

rcollyer

@OleksandrR but, the president/CEO of a company can be sued in civil court as part of the action against the corp.

J. M.

@rcollyer I distinctly recall a few pharma-related tort where exactly that was done...

rcollyer

4:38 PM

@OleksandrR I am not advocating getting rid of the idea of a "legal person," but I don't think they should have the same rights as an individual.

J. M.

A subset, maybe, but certainly not the whole bunch.

rcollyer

Nor, do I think they should have rights, per se, at all, but privileges.

Oleksandr R.

@rcollyer okay, well that's fair enough, and I do agree with you. But the common conception of "an individual" is legally known as "a natural person". There are other types of persons too, that have different types of rights.

rcollyer

@OleksandrR true. Let me clarify that then. A legal person (as distinct from a natural one) should not have any rights, per se, but should be granted privileges which may include some of the rights of a natural person. This includes governmental bodies. A governmental body, though, would have access to some privileges that should not be granted to a natural person: removal of property, liberty, or life come to mind. A non-governmental legal person, though, should never approach such privileges.

Oleksandr R.

@rcollyer a very interesting point in itself. Should natural persons have rights (say, constitutionally enumerated), then, or are there no "rights" per se but only laws governing how persons behave toward one another? (I would be in favour of the latter, since it is more legally actionable.)

rcollyer

4:47 PM

@OleksandrR the latter contains a trap: by restricting a natural person's rights to an enumerated set, then any extension of those requires extraordinary action, hence the inclusion of the tenth amendment in the US Bill of Rights. It implies that non-enumerated rights still exist. This was used in Roe v. Wade to imply the right to privacy.

Oleksandr R.

@rcollyer a good point, although then we have the difficulty of determing where rights come from in the first place. Let's take, say, the right to life, liberty, and pursuit of happiness. Legally, how is "pursuit of happiness" (or even happiness itself) defined and who has responsibility for protecting this right in practice?

@rcollyer I think the 10th amendment goes a long way to avoiding that specific problem by saying that rights (of the Federal Govt) have to be enumerated and are only legally actionable as such.

rcollyer

@OleksandrR not an unreasonable point. To me, the amendments read as if they come from experience with a corrupt government, one that will suppress and oppress the people to remain in power.

And a gov't bound by its own laws is responsible for ensuring that the people retain access to their rights.

Oleksandr R.

@rcollyer yes, I agree. Generally they are negative rights (prohibiting unwarranted interference), rather than positive ones (guaranteeing certain opportunities). I think the former is less legally troublesome.

rcollyer

@OleksandrR There remains wiggle room, though, in the elastic clause.

@OleksandrR exactly. That way they remain enforceable.

J. M.

@rcollyer Huh, I haven't heard of that before. Thanks for that.

rcollyer

4:57 PM

@JM My mother-in-law is a middle school social studies teacher, so discussions about civics is fairly common.

Oleksandr R.

@rcollyer indeed. I think US constitutional law is a fascinating subject. In particular, sometimes the reasons for the Supreme Court deciding in a particular way seem quite unexpected (for example in Brown v. Board of Education, where the decision was undoubtedly the morally correct one, but purely legally that interpretation doesn't obviously suggest itself, at least not to me).

rcollyer

@OleksandrR that is an interesting point. A previous court ruled that "separate but equal" was fine, but in Brown v. Board, though, it was argued that "separate" implies "unequal" violating equal protection under the law.

Oleksandr R.

@rcollyer yes, exactly. Undoubtedly in practice the separate facilities were not at all equal, so Brown could have won his case on that point alone. The aspect of it I find particularly interesting is the precedent that "separate" was thereafter considered to be inherently unequal.

rcollyer

@OleksandrR I find this map describing the segregation particularly interesting.

Is this question a candidate for migration?

Oleksandr R.

Another clause I find has an interesting application is the regulation of inter-state commerce. The ways in which certain sorts of activity are designated commerce or not, or inter-state or not, are fascinating to read about.

rcollyer

5:10 PM

@OleksandrR true. very true.

Pretty much anything that crosses a state border falls into Federal jurisdiction, that they can take notice of if they wish.

How about this question? Would it, also, be a candidate for migration?

J. M.

@rcollyer Yes, both look good.

Oleksandr R.

@rcollyer no, it's broader than that. For instance I recall a case of a farmer who grew and milled his own wheat in order to avoid having to buy it from another state which at the time was subject to quota. (Unfortunately I forget the name of that case.) The ruling was that his activity could be regulated simply because it was undertaken in order to avoid inter-state commerce.

rcollyer

@OleksandrR A case that was merged into the Brown v. Board case at the Supreme Court hearing was the Davis v. Board case where the students staged a walk out due to the terrible condition of their school.

@JM I flagged them both using "This is borderline off-topic here, and will likely get better reception over at Mathematica.se. It should be migrated." which is less true on the second question, but we'll see.

I don't have a high flag weight, so it may be a couple of days.

@OleksandrR I vaguely remember that. I'm not sure I agreed with the decision.

At 310 consecutive days on SO.

Oleksandr R.

@rcollyer I also find the case of the sugar growing industry rather fascinating and perhaps amusing. As a result of extensive laws basically propping up sugar prices enacted in the name of regulating inter-state commerce, the market for sugar in the US was almost destroyed and the situation persists to this day that high-fructose corn syrup is extensively used instead of sugar in the US but is almost unknown anywhere else in the world.

rcollyer

@OleksandrR I was not aware of that.

Heike

5:23 PM

@OleksandrR I thought that corn syrup was a result of subsidised corn production.

rcollyer

@OleksandrR Sometimes regulation is a good thing (think: banking industry), sometimes they don't make nearly as much sense.

Oleksandr R.

@Heike that too, yes--regulation on both sides contributed to the situation.

J. M.

@OleksandrR The upshot of that is that HFCS is even more likely to trigger diabetes in sensitive individuals than sugar itself. Oh well.

rcollyer

Suggestions on the tag for this one:

0

Q: Defining Tags in Reap (Mathematica)

I am trying to use Sow/Reap in Mathematica to replace Append in my code. The problem is that I need to define the tags within my code, but doing so doesn't reap the desired results. A simplified version of my code is Clear[tag]; Reap[tag = {a,b}; Sow[1, a];Sow[2,b];, tag] However, if I define ...

untagged

@JM I really appreciate your licensing terms. :)

Heike

@rcollyer functions maybe?

rcollyer

5:26 PM

@Heike it was a thought. Do we need a special one for Reap and Sow?

J. M.

@rcollyer Oh, the WTFPL? I wouldn't be posting code I didn't want other people to use anyway... so it seemed natural.

Heike

@rcollyer This question had more to do with the HoldFirst attribute than with Reap/Sow itself.

rcollyer

@JM I understand that. But, it amuses me, nonetheless.

@Heike reasonable. Edited.

Oleksandr R.

@rcollyer I agree, albeit I would personally distinguish more strongly between helpful and unhelpful regulations rather than industries requiring or not requiring regulation. I think some of the regulations placed on the banking industry have been net harmful by contributing to the formation of just a few very large banks in preference to a plurality of smaller ones in competition.

J. M.

@OleksandrR ...and you have some of these large banks buying up the small banks every so often...

rcollyer

5:31 PM

@OleksandrR I was under the impression that it was deregulation that precipitated those mergers.

@JM the op on the VBA question is requesting a delete. I'm inclined to keep it, though, as it may (eventually) be solved. Thoughts?

Oleksandr R.

@rcollyer deregulation certainly allowed the mergers to happen, but also to a large extent the legal burdens now placed upon banks (under, e.g., the PATRIOT act) allow only those big enough to afford a large team of well paid lawyers to survive.

rcollyer

@OleksandrR that is true. that law should never have been written.

J. M.

@rcollyer I'd rather it wasn't deleted.

PATRIOT was a kneejerk reaction, plain and simple.

rcollyer

@JM my thoughts, too. It isn't a bad question, just very ambitious.

@JM yes it was.

J. M.

My question a few lines ago still stands. I'd love to see someone try to reimplement FindClusters[]in VBA...

rcollyer

5:39 PM

I think GatherBy would require a bit of work, by itself, but would not be the tour de force that FindClusters would be.

Oleksandr R.

@JM I don't know. Wouldn't be too difficult to implement K-means, I think? Whether it's worth doing in a language like VBA is another matter, of course.

rcollyer

@OleksandrR that is a good point.

I have spent more than enough time here today. Bye all.

Heike

bye

Oleksandr R.

Bye. Nice talking!

I wonder if the VBA stuff could perhaps be attacked by way of F#, which would make the pattern matching aspect a lot easier. (I don't actually know F#, so how feasible any of it would be even then, I'm not sure. But at least you wouldn't have to fight VBA.)

J. M.

@OleksandrR Maybe things are better these days, but I do recall the method being a bit elaborate...

Oleksandr R.

5:54 PM

@rcollyer I know you've gone now but I thought you might appreciate this: some researchers in the European Psi-K collaboration are having a density functional popularity contest! If you like or hate any specific functionals, you can cast your votes now. :)

J. M.

@OleksandrR Which one's leading?

Oleksandr R.

@JM I think K-means isn't very hard although it's not necessarily that effective. Better methods like kernel PCA, though, I definitely wouldn't want to try in VBA.

J. M.

@OleksandrR Ah, well, SVD (PCA's engine) will definitely be a bitch to do in VBA... :D

Oleksandr R.

@JM No clue. I just got this in my email this morning. These are solid-state physicists though so I think anything B3LYP related is going to crash and burn.

@JM unrelated, but do you know of any ICA implementations in Mathematica? For a while I've felt it could be useful but I don't know that much about it and don't really have time lately to read the literature and implement something myself.

J. M.

@OleksandrR In Mathematica, no. I would admit I haven't gone through the papers for those in detail, though, nor have I extensively experimented with the methods, so I can't give a "X is the best" opinion. Sorry. :(

...though translating this to Mathematica seems to be not too difficult to do.

Oleksandr R.

6:06 PM

@JM not to worry. I did a quick search for Mathematica implementations previously but it seems like nearly everyone uses MATLAB. Unfortunately I don't know MATLAB and since I also don't understand the methods, translation is not exactly feasible. Oh well...

J. M.

(from what I've seen in the code)

Oleksandr R.

Yes, I did see that one, thanks. I think I would have to start from the Python implementation personally, but I agree, translation should be possible.

Seemed a bit surprising that nobody's put anything together in Mathematica yet, though. On the other hand I had to code Savitzky-Golay myself, too...

J. M.

@OleksandrR Heh, I managed to derive my own implementation of Savitzky-Golay... ah, memories.

Were you using an LU decomposition of the design matrix, or something else?

Oleksandr R.

@JM for S-G? I just used PseudoInverse. But there's the undocumented LinearAlgebra`VandermondeInverse too, actually.

J. M.

@OleksandrR Ah, $O(n^3)$ (PseudoInverse[] that is). The one I wrote was a bit more elaborate, but $O(n^2)$. I'm not sure what LinearAlgebra`VandermondeInverse is using internally, but I suppose it's the $O(n^2)$ one...

image_doctor

6:13 PM

Hithere, Would anyone know how I can clear the value Out ?

Szabolcs

@imagedoctor Use the CleanSlate package's ClearInOut function or Unprotect and Clear the Out symbol.

J. M.

And I only needed a one-dimensional array for storage.

image_doctor

@Szabolcs thanks that gets me 7GB back :)

J. M.

@image_doctor o_O what have you been doing that has you obtaining output in gigabytes? :o

Oleksandr R.

@JM yes, I wasn't too concerned with that since the design matrix isn't really large, and anyway it seemed to me that it was sometimes better to do the inverse in exact rationals due to numerical instability for high-order polynomials and long windows.

@JM also I needed the whole matrix to get the skewed windows needed to deal with the endpoints.

J. M.

6:18 PM

@OleksandrR Heh, the method I came up with also did special treatment for endpoints. But yes, if you just need something to get it done quickly, it's a good standing start.

(Maybe I should post about it someday. It was pretty clever math buried within the pages of an analytical chemistry journal...)

Last, I checked, a lot of people use something similar to the implementation in Numerical Recipes, and that demands a scratch two-dimensional array. Eep.

Heike

The combination of the terms "large engineering department" and "excel spreadsheets" in this question worries me somehow.

4

belisarius

Hey! 5K :)

mathematica.stackexchange.com/users/193/…

Heike

@belisarius congrats

belisarius

@Heike Thanks!

Oleksandr R.

@JM I'm not at all sure that my method deals with the endpoints correctly when you're taking the derivatives rather than just smoothing, so I would definitely be interested in seeing that. (I never needed the derivatives; smoothing was enough for me, so I didn't really go over it too carefully in that case.)

J. M.

6:22 PM

@OleksandrR Oh, I covered derivatives too. :) Okay, I'll write a blog post on it sometime this week...

@belisarius 'grats!

Oleksandr R.

@JM that's very good of you. Thanks!

belisarius

@JM Bow!

J. M.

@Heike Notwithstanding the "Excel for X Engineering" titles that keep cropping up every few years? :D

@OleksandrR (It's about time anyway, I've been sitting on my algorithm for too long and I weep every time people point to the NR implementation...)

rcollyer

@OleksandrR well, that let's me know how much I don't know about the functionals out there. :P

Oleksandr R.

@JM I hear you on that.

@Heike The mere mention of "Excel spreadsheets" worries me every time. Recently a colleague confessed that he uses Excel for his data analysis and even production of graphs for publication. Needless to say I was shocked.

J. M.

6:26 PM

So, how come almost every engineering library I've seen keeps a copy or two of those "Excel for X" titles? :D

Heike

@OleksandrR I have a friend who wrote software to check the validity of excel code. According to her a lot of companies use excel for their administration but are reluctant to admit it.

J. M.

On second thought... I'm not sure I'd like to hear the answer.

Oleksandr R.

@Heike I'm reminded of this: digilander.libero.it/foxes/MultiPrecision.htm. While a valiant effort to make Excel useful for something, development was halted because Excel's just so broken that even using extended precision isn't particularly effective for helping it get the right answer. Not encouraging for financial applications!

Heike

@JM Doorstops? Or maybe to prop up the computer screen.

J. M.

6:40 PM

@Oleksandr: if you'd like to experiment with at least the method for generating the smoothing coefficients now (as opposed to the derivatives), here it is:

gramP[k_, m_, t_] := (-1)^k HypergeometricPFQ[{-k, 1 + k, -m - t}, {1, -2 m}, 1]

With[{m = 5, n = 2}, Table[Sum[(((2 k + 1) (-1)^k Pochhammer[-2 m, k])/Pochhammer[2 m + 1, k + 1]) gramP[k, m, i] gramP[k, m, t], {k, 0, n}], {t, -m, m}, {i, -m, m}]]

(That's a toy implementation at least; the actual one uses recursions. That I'll post when I get to writing the blog entry.)

Oleksandr R.

@JM that is very nice. :)

R.M

J. M.

@OleksandrR Just for clarity: n is the order; 2m + 1 is the number of points. The middle list is the centered window, and the others are for the endpoints.

Oleksandr R.

@JM yeah, I understood. Unsurprisingly the result is identical to what I got myself. The method is obviously substantially different, though.

Heike

@RM makes me wonder what the original books were.

Szabolcs

6:48 PM

This question is asking about how to prevent Eigenvalues from "reordering" the eigevnalues (sorting by magnitude), but unless I'm mistaken there's no order defined by the matrix, so it's not clear to me what the OP means.

Oleksandr R.

@Heike "Microsoft Calculator for Nuclear Reactor Engineering"?

Heike

@Szabolcs I was in the middle of asking the exact same thing when your comment to the question popped up.

image_doctor

@JM I'm not entirely sure :S

J. M.

@Szabolcs No idea either. Why not just use SortBy[] after applying Transpose[] to Eigensystem[]'s output?

R.M

@Heike I took the image from this thread... I was originally going to write over all the titles, but then I realized it's hard to do write in small letters with my finger and a trackpad

Heike

6:52 PM

If you take for example ll = {{1, 0, 0}, {0, 2, 0}, {0, 0, 3} and qq = {{0, 1, 0}, {1, 0, 0}, {0, 0, 1}} then qq.ll.qq^T = {{2, 0, 0}, {0, 1, 0}, {0, 0, 3}} so there is no order of eigenvalues as far as I can tell.

@OleksandrR "Excel for nuclear engineers"

J. M.

@Heike Precisely. If you want your eigenpairs to be sorted according to some criterion of yours, that's what SortBy[] can be used for...

@Heike Horrors.

Sjoerd C. de Vries

Although I find the question interesting I wonder whether this doesn't precisely matches the 'not constructive' question category. What do you folks think?

11

Q: Suitability of Mathematica as platform for engineering calculations and programs

I am considering using Mathematica with Wolfram Workbench as a standard platform for calculations and programs in a large engineering department. I am looking for a solution that would provide better validation, documentation and revision control than Excel spreadsheets with VisualBasic macros. A...

programming

J. M.

@SjoerdCdeVries I would say it's an experiment they really need to perform themselves, but maybe somebody who went through something similar (migrating from systm Z to Mathematica) might be able to provide a usable answer...

Sjoerd C. de Vries

@j.m. True, but is this a type of question we want here?

J. M.

It is admittedly soft, but I don't feel too strongly about it to want to close it.

Heike

7:01 PM

@SjoerdCdeVries Given the large number of upvotes in so short a time quite some people seem to find it an interesting question.

Sjoerd C. de Vries

@heike I thought it interesting too, but it isn't really complying to the rules as it solicits opinions.

J. M.

@SjoerdCdeVries Too soft, huh? Maybe we should let users cast the close votes for this case...

Heike

@SjoerdCdeVries That's true, although the reactions so far seem to be quite objective.

J. M.

@Heike That too. Leonid's reply is quite pragmatic...

Sjoerd C. de Vries

It's not that I don't like the question, nor Leonid's answer. It's that you lose the moral right to close any other question for being 'not constructive' after letting this go.

R.M

7:17 PM

@SjoerdCdeVries I felt that question was just as subjective as this one... The reason I didn't vote to close either of them was because although they were soft/subjective, both were well reasoned, coherent, gave a practical situation that they faced and I got the feeling that they (at least the OP of this new question) had a grasp of what they would like to take away from the comments/answers.

It wasn't a blanket "Has anyone developed a large system in mma and if so what are the challenges you faced?" kind of question. Perhaps, this might be better suited to Programmers, but I wouldn't be opposed to leaving it here (would argue for it to be here, in fact).

I don't think that by not closing this, we have lost any moral ground in closing other NC questions.

Sjoerd C. de Vries

@r.m ok, I can live with that reasoning

Oleksandr R.

@SjoerdCdeVries aren't we allowed to have subjective questions? I realise they're not welcomed on some other sites, such as SO, but this being our site, can't we decide for ourselves whether it's a good question or not? IMO, asking for opinions in the context of a community of experts is different to standing on street-corners shouting for help (as one might interpret such a question on SO, especially e.g. C++ or Java tags).

Szabolcs

7:35 PM

I still don't understand the eigenvalue ordering question.

how does any basis define the ordering of the eigenvalues?

@RM "Are the users scientists or are they knowledgable in programming and software development?" <-- so are those exclusive? :P

R.M

Most certainly :)

Szabolcs

@Sjoerd Did you really understand that eigenvalue question?

Heike

7:50 PM

Why is rice pudding pie so addictive?

Oleksandr R.

8:04 PM

@Heike I would never have thought to put rice pudding into a pie! But it sounds like it would be healthier than, say, a cheesecake, which would be an advantage if it's addictive... might have to try that one.

Okay, so everyone's heard of the Jacobian and the Hessian... does anyone know if there's a standard name for the tensor of third derivatives? If there is, I haven't heard of it.

Heike

@OleksandrR This type of pie is typical for the south of the netherlands and parts of belgium. The crust is made of a yeast dough and can have various fillings such as cherries and apricots, but my favourite by far is the one with a rice pudding filling, especially when topped with whipped cream and chocolate shavings.

@OleksandrR Maybe you can call it the Oleksandrian

Oleksandr R.

8:20 PM

@Heike I think that would be tantamount to "Tai's model". ;)

Heike

@OleksandrR At least you'll be famous.

R.M

@Heike I'm sure that sentence is missing an 'in' somewhere ;)

Sjoerd C. de Vries

9:48 PM

@OleksandrR Sure, we can have subjective questions, but not any subjective question. Differentiating between the good and the bad ones is, unfortunately, a bit subjective. Always good to discuss them to see what fits this site. I found the following blog useful in this regard.

Blog - Stack Exchange: Good Subjective, Bad Subjective

Robert Cartaino on September 29, 2010

Stack Exchange is about questions with objective, factual answers. We’ve been crystal clear about this for as long as I can remember, even back to the earliest, pre-beta days of Stack Overflow. It’s right there in the standard Stack Exchange FAQ:

What kind of questions should I not ask here?

Avoid asking questions that are subjective, argumentative, or require extended discussion. This is not a discussion board, this is a place for questions that can be answered!

Thus, questions that are not answerable — discussions, debates, opinions — should be closed as subjective. It seems simple enough: Fact good; opinion and discussion bad. But why? …

Compare this with the description of the 'not constructive' description "As it currently stands, this question is not a good fit for our Q&A format. We expect answers to be supported by facts, references, or specific expertise, but this question will likely solicit debate, arguments, polling, or extended discussion. If you feel that this question can be improved and possibly reopened, see the FAQ for guidance."

@Szabolcs No, not really. This is not an area in which I feel myself particularly comfortable.

Transcript for