Why Mathematica.SE and not ComputerAlgebra.SE?

Question

Why at the creation of Mathematica.SE the decision was taken to focus only on Mathematica rather than the whole field of symbolic computation ?

The beta seems to be successful so I can't really be critical, but I can't explain it either. Here are my arguments in favor of the existence of SymbolicComputation.SE (lets call it SymbComb.SE, this not a name proposal, just a variable name) :

• Mathematica is not prominent, neither in research nor in education ;
• some computer algebra systems are at least as important (both is user mass and scientific impact) as Mathematica is, I think of Maple, Sage and Magma ;
• some computer algebra systems are not as rich as Mathematica but they are irreplaceable in some specific fields : commutative algebra has Macaulay2, Cocoa and Singular, group theory has GAP, number theory has PARI/GP, etc ; and all what is done with these systems cannot be done with Mathematica ;
• most of (researcher) users of Mathematica are competent with at least one or two others comp. alg. system, so firstly a hypothetic SymbComp.SE would not be split into different communities and secondly Mathematica.SE does not exploit the full potential of the contributors ;
• SymbComp.SE would gather much more people than just Mathematica.SE, but would still be under the critical point where bigger is not better ;
• Wolfram is insane.

What do you think of all that ?

---

EDIT

Well, thanks for your answers, this is more than expected !

To downvoters, this has never been a proposal of any kind concerning the future of mathematica.SE, have no fear I don't want to steal your good.

I wanted to understand the—apparent—paradox that there exists a rather successful Mathematica forum but no computer algebra one, whereas it seemed to me the latter should have wider audience. Thanks to your answers I do understand. I haven't realized that there is a united (and a bit touchy maybe ;)) community gathered around Mathematica, and for an important part more interested in the (rather awesome) language, the (very good) interface or other original features than in doing mathematics and a fortiori computer algebra. I haven't realized it sooner because in research, there is a computer algebra community, centered on what we can do with computers, and not really communities centered on specific softwares — except maybe some open source scholar soft. At least, that's my feeling.

I hope you won't stone me if I dare to ask a question on the forum ;) I might feel the need for it : I do use Mathematica for my work !

Answer 1

I think the simplest answer is, because those of us who started this were interested in mathematica and not (say) maple. That's all there is to it.

I can say for myself, for instance, that I am not interested in computer algebra, and don't use mathematica for computer algebra, apart from checking integrals once in a while; I use it for numerical computation (I am just offering this as an example, not claiming everybody is like this) and generally to play with algorithms before writing them elsewhere if I need to. I don't even think of it as a computer algebra system; I think of it as a language with a huge library of things I don't need to program (like Python, but higher level and with more useful stuff for me).

So basically we're not people doing computer algebra who happen to use Mathematica. We are people working on various thigns, who happen to use Mathematica for work, fun or both. So in that context it makes perfect sense to form a community around it, since we can answer each other's questions.

As for the downvotes, here they simply indicate disagreement (rather than that there is something wrong with your question, as they would on the main site).

Answer 2

I think that this community never aimed to "gather much more people than just Mathematica.SE", as that would mean that we have to deal with questions concerning Maple, Matlab, Octave, Sage, and many other CAS. This flood of diverse questions would fragment this community, which shares one rather important feature: we all use Mathematica for work (or hobby). I myself am not particularly interested in other programming languages at the moment, and even when I was, I could easily find other dedicated sites for e.g. Matlab or R.

I don't think that any of the languages you have listed have anything that "cannot be done with Mathematica". These are all Turing-complete languages, so you might only say, there are languages that are better fit for certain tasks. While of course each of your points can be countered individually following the above lines, I don't think this would be a creative answer. Also, I have no intention to prevent any SymbComp.SE to bootstrap itself. I merely list some points that were considered when this site was branched from StackExchange.

I think I can say on behalf of the community that we do believe that both Mathematica.SE and Mathematica itself are still gaining momentum day by day, year by year, extending over more and more domains and problems. Statistics show that people interested in Mathematica are gathering here with a steady pace. I'm pretty sure that WRI has also statistics that industry, education, academia, etc. are attracted more and more to their products (at least their development pipeline indicates this, e.g. all the curated data, Wolfram Alpha, SystemModeler, etc.). Accordingly, we believe that Mathematica is getting better and better, and we think that we will use it for more and more tasks (as it happened to me at least in the last 10 years). Consequently we want to invest more into understanding the subtleties of the system, which can be done the easiest if there is nothing else distracting. As an example, I think that mixing different languages on the same site is not helpful or productive but rather confusing and - to be honest - time wasting. Of course tags could be used to filter unwanted questions, but frankly, I like the fact that I did not have to filter out anything so far on Mathematica.SE.

To give an even more general answer: While there are many interesting addictive things out there, we are the sick people gathered here to overcome our issues in the Mathematica therapy group.

Answer 3

Mathematica.SE was initiated by a group of people who gathered under StackOverflow's mathematica tag. This tag existed already for a few years but really took of about a year ago. We encountered lots of difficulties from people that didn't know Mathematica at all but considered many Mathematica related posts off-topic and consequently closed them. This was one of the reasons to go for a dedicated mathematica site through the area 51 process. So basically what we have here, is our old mathematica tag with more freedom to discuss issues we find relevant. For most of us that's enough.

Apart from Matlab, which has a large tag on SO, the other systems you mention (Maple, Sage and Magma) lead a marginal existence on SO. Not much synergy to be gained from including them. There would also be the problem of flame wars or other unpleasantries that we would like to avoid.

Answer 4

At the risk of piling on, the other answerers are correct in stating that merging Mathematica into a "computer algebra" site would severely limit the scope, compared to what Mathematica is capable of. Mathematica is not a computer algebra package. It goes far beyond that, and many of us barely use its algebra capabilities explicitly. I am an economist working outside academia and mainly use it for numerical computations and data visualisation. Mathematica has a much wider audience than researchers in computer algebra.

All of this was hashed out in Area 51 last year, but unfortunately most of the discussion was linked to the previous Mathematica proposal that was closed, not the one that got up, so we can't link to it.

As for your dot-point assertions, I would suggest that:

1. is debateable (relative to the CASs you mention, it is false in private industry and government);
2. doesn't mean we have to share a site even if it is true, and I suspect it is untrue in private industry and government;
3. may or may not be true but, again, users of those systems can set up their own site if they want;
4. is demonstrably untrue, especially outside of academia;
5. is true in the sense there would be more users, but as explained in other answers, would be no advantage to those users;
6. is irrelevant.

Please do not take our responses as anything less than welcoming to the site. If you have Mathematica questions, we would love to see them on the site! We were the least-downvoting tag on StackOverflow back when we were there, and we try hard to be a welcoming community that encourages newcomers. If we sound touchy it is because the first Area 51 proposal was closed by StackExchange management, and across the two proposals, we have had to hash out this question several times. We have also had to deal with the outright hostility of those who think that anything that is not open-source is somehow suspect and to be opposed.

Answer 5

I certainly agree with the comments of Istvan, RM, and Sjoerd. In addition, I'd say that discussion about Mathematica covers much more than computer algebra. This group has hundreds of tags covering numerics, graphics, interactive dynamics, functional programming, optimization, linear algebra, typesetting, import/export, mathlink/jlink, topical areas (physics, chemistry, math), the list goes on.

Conversely, the specific algorithms of computer algebra are actually discussed very little. A search for Groebner returns 3 hits, for example. The group you propose might be interesting, but I think it is very different from this group.

Answer 6

SE sites are aimed to be built around communities, not ideas. Usually, the typical Area51 process involves someone pitching a random idea, with others joining if they liked it. It trudges along through the process and eventually enters beta... and then what? There's no community (yet), which means everyone has to find common middle ground, argue about Maple vs Mathematica, bicker about open source vs closed... it'll be chaotic, just like your average CAS usenet group.

In this case, there was an existing community on Stack Overflow (SO) that had a distinct culture that clashed with the rest of SO. Also, the scope of SO severely clamped down on the type of questions that could be asked, leading to the creation of this site. That is not to say that there was no drama or that creating a super-CAS site wasn't suggested (I can't find this discussion, as it was removed when the old Mathematica proposal was deleted).

Note that none of the other systems you mention had any semblance of activity, let alone a community when compared to Mathematica (cf. maple, maxima, sage, mathematica). Who is around to answer questions on these topics? If this was billed as a generic CAS site and Maple questions routinely go unanswered, then we get a bad name. Besides, I'll consider Maple/maxima/sage to be on the same footing as Mathematica when you can write interactive books and create interactive presentations with them.

Answer 7

there is Computational Science SE Beta that is definitely geared toward computational science