24

Wolfram has been making some big announcements lately regarding Wolfram Language which seems--for the most part--to be Mathematica-"language" accessible from other interfaces. Browsing through the documentation, I see some new functions and ideas (ex: Association) that we currently don't have. What are the plans for Mathematica SE if we start seeing the traditional Mathematica language in other areas?

19

There are three separate but interrelated questions here, touched on by the other answers and the many comments.

  1. Should we treat questions about the Wolfram Language as on-topic even when it is not specifically about Mathematica, the product? (And if so, does this change our decision to make Wolfram Alpha questions off-topic?)
  2. Are these new Wolfram Language products, such as the recently announced implementation on Raspberry Pi, an opportunity to promote Mathematica.SE?
  3. What should be the role of the Mathematica.SE community in building a broader following for both Mathematica the product and the Wolfram Language more broadly?

My own answers to the above three questions are “yes”, “yes”, and “possibly a bit more than we are currently doing”.

My rationale for answering “yes” to question 1 is that the Wolfram Language implementation is essentially equivalent to the kernel in the Mathematica application, or maybe a future version of it. There are also questions about the Mathematica front end that are explicitly on-topic (and were a lot of the reason why the Mathematica community could never have stayed on StackOverflow). There is a question about whether other “front ends” are on-topic; I don’t have a strong view on this.

On question 2, I think we are already getting an uptick in interest in the site because of the Raspberry Pi publicity. We could find a way to capitalise on this.

I sympathise with the frustration about the way Wolfram Inc markets its products. You only get to cry “New Kind of Breakthrough” so many times before people stop listening. In many ways, this site could serve as a source of less-hyperbolic advocacy, making the practical case for Mathematica the product and the related Wolfram technologies. How exactly we go about that, if indeed that is what people want, is a matter for a separate meta thread and/or chat event, I think.

  • 2
    I fully agree with you Verbeia, but I do not have time right now to post a more in-depth answer. To all others worrying about this and that, I only wish to say: relax and enjoy the ride. S.W. is not only a great scientist, but has also shown in the last 25 years that he is a wise CEO. WRI, needs to find its killer product/market like Apple with the ipod/itunes did. WA first and embedded WL now might be it. – magma Nov 24 '13 at 10:54
  • 5
    I too fully agree with all points in this answer. I would welcome questions about Wolfram Language on or off the RPi (perhaps even include it in our FAQ/rules), but I am strongly opposed to renaming the site WolframLanguage.SE or any such variant of that (there were a few questions on this in chat). – rm -rf Nov 24 '13 at 18:08
  • 3
    @rm-rf I wonder of the logo might be changed to Mathematica with "and the Wolfram Language" as a subtitle? Might also stop some of the math homework, too. At the very least we should consider changing the FAQ. – Verbeia Nov 24 '13 at 18:57
  • 2
    Agree with your points. Re 3, it would also be nice if we could get back some of the friendly welcome spirit that used to characterize the site in the early days. The WL will attract a more diverse audience, and we might need to exercise patience... :) – cormullion Nov 24 '13 at 20:08
  • broadly agree. 2 cents: I think how it markets its products is less of a problem to how it develops its products. – Mike Honeychurch Nov 24 '13 at 20:12
  • 6
    I totally agree. Up until now it has been W|A that attracted the widest interest, but its seemingly arbitrary input interpretation and largely unknown capabilities make it very difficult to have any constructive questions about it. I still consider this to apply and so do not support allowing W|A questions. The Wolfram Language should (hopefully) be much more predictable and in many ways is a known quantity. I'm glad that this might potentially become most users' first point of contact with Mathematica and related technologies, rather than W|A, and would be happy to see questions about it. – Oleksandr R. Nov 25 '13 at 1:19
  • Very well summarized, and I also agree with your answers. – Szabolcs Nov 25 '13 at 21:37
17

I agree that it's a "New kind of Marketing" for Mathematica, but in my opinion its a important steep towards make Mathematica more known in the traditional programming community. Today, I believe that 99,9% of programmers think in Mathematica (when they know it exists!) as an tool for Physicist and Matematicians, and Wolfram wants to break this concept, making it an really language. I like to call it W, just as we have R. :)

When I see Tiobe list today, I get worried about the future of Mathematica, and I fell Mathematica really needs a new strategy to reverse that. I agree that language rename is a nice try.

I have send an mail to Tiobe, here is my question:

Hi

I would like to know if Wolfram Mathematica could not participate. I think that with wolfram alpha it gets a little bit more popular no? It has a very active google groups forum and exists since 1988.

Tks in advance Murta

and the answer:

Hi Rodrigo,

Thanks for your feedback to our TIOBE index. Mathematica has been tracked for almost 10 years now. Currently Mathematica is at position 128 down from position 118 last year. In the past it was quite popular but as you can tell from the attached trend graph its popularity is declining. I hope that this answers your question.

Regards, Paul

...with Mathematica chart attached.

enter image description here

Let's see what the future bring to us.

  • 5
    Though I also want Mathematica/WL to be successful I am not sure this push is going to be the way to turn the Tiobe numbers around. Notice Matlab is 16 ... and it is un-apologetically an engineering language, similarly R is 31, another special purpose language. I worry that focusing on embedded systems and other areas will only weaken Mathematica's strength in science (for example I would rather have robust panel plot alignment than rasberry pi support, cool as it is). Trying to make WL compete with Python etc seems to ignore what Mathematica is strong at ... ie science/math. – Gabriel Nov 22 '13 at 13:45
  • 2
    @Gabriel I think it's really great to have embedded systems support. I want this for such a looong time, yet I'm totally not familiar with hardware development. I always consider MMA not only a scientific computing tool but also a full-featured program language. But Because of many reasons (the price, the steep learning curve, etc), it has been in the ivory tower since ever beginning, away from the daily life where it should have been able to show its capabilities. The free port on RPi (and maybe more platforms) makes things different. I think it's a potential marketing strategy, – Silvia Nov 22 '13 at 14:55
  • 2
    ... at least here in China, where many hackers are not willing to buy software but very willing to buy chips and boards. btw I myself bought my first RPi 2 hours after reading the blog. Hope it a pleasure experience. :) – Silvia Nov 22 '13 at 14:56
  • 3
    I think the rasberry pi support is awesome ... but what does WL give you that python doesn't? Outside of making it easier for people that now use WL? I am not saying that this development is bad, but I don't believe it will make a difference for the Tiobe numbers. My friends checked this out for a second when I sent the link, then just went back to python (and proceeded to make fun of me for the crazy way that S. Wolfram writes his blogs ... so many breakthroughs and revolutions ...) – Gabriel Nov 22 '13 at 15:00
  • I also fear that we need to rebrand CDF to WDF, hell even mathlink is getting a wolfram ... – Gabriel Nov 22 '13 at 15:07
  • @Gabriel As I understand, "WL" on the RPi is the full Mathematica kernel. It gives you all the builtin functionality of the kernel. To call it a "language" is very misleading and misguided marketing; this seems to be just the kernel with full functionality, and not the pure language with libraries removed. So comparison to Python is not appropriate here. Comparison to a set of Python libraries and their (non-)integration would be more appropriate. – Szabolcs Nov 22 '13 at 15:30
  • 5
    I don't think traditional programmers will find Mathematica useful for their work. Most of them won't even "get" it---just look how long it took for the notebook paradigm to be adopted by others (see Sage). Mathematica shines when used interactively, and IMO it's considerably better in that area than Python because Python just wasn't designed for interactive work. Even when people use Python for data analysis or other interactive work, the typical workflow is "write script first, run later", which is not really interactive ... But Python is a great general purpose language which M isn't. – Szabolcs Nov 22 '13 at 15:35
  • 1
    Maybe, this might be from the matlab-inspired side of the scientific python group (numpy/scipy). The ipython and Sage groups where very aware of the need for Notebooks, they just had to code if from scratch. Anyway I don't think this matters for the Tiobe point ;) I love Mathematica, I just worry about Wolfram loosing its audience adding to many "general" features without giving better science/math features (like DataFrame like objects ... I am so scared that we just get associative arrays -- even matlab has dataframes now, will suck to be the only science language without this feature) – Gabriel Nov 22 '13 at 15:58
  • 1
    @Silvia I think we all agree these new features are cool, I just worry that they come at the cost of more scientific features (like DataFrames and Plot Alignment etc). I also am going to get a RPi to play with this. But for my day to day work I would rather Mathematica was going the route of matlab/R as becoming a better platform for doing science specifically. – Gabriel Nov 22 '13 at 17:18
  • 5
    I am trying understand why Tiobe even matters. According to its site "The ratings are based on the number of skilled engineers world-wide, courses and third party vendors." So if you use the language as part of your work but you aren't a full-time programmer, if you don't need a bogus accreditation course (and anyway you can learn yourself through existing resources) and if you don't need third-party vendors because so much comes in the box - of course that language will score lowly. – Verbeia Nov 22 '13 at 22:29
  • 1
    (continued) Also, that graph doesn't mean that Mathematica's absolute popularity is declining, only its relative popularity, and that doesn't mean it isn't successful and influential. Yes, there is a lot of boring corporate development happening in various Microsoft languages. But if that development is highly inefficient, needing lots of process and many developers, then yes, the Tiobe numbers will look good for those languages. For science/finance/technical users who aren't professional programmers, Mathematica is a better option. – Verbeia Nov 22 '13 at 22:32
  • 7
    @Verbeia I think this is not so simple. For the language to be widely successful, it needs a number of things. It needs a strong developer community (not just user community), it needs effective means to educate newcomers to the language (and no, self-learning using Help etc is not it - it is a luxury which not many can afford. If the company decided to adopt the technology, it needs ways to jump-start its employees on it quickly). It is also the job market, availability of positions, etc. If all of these are lacking, all we can talk about is more or less niche adoption. – Leonid Shifrin Nov 23 '13 at 13:43
  • 9
    @Verbeia I think this is not necessarily bad, there are many cool languages which are very effective in their primary fields of application but very little known to the wider programming audience, but these languages do not have a massive impact on what is currently being used for development in the majority of companies - and I think, Tiobe is measuring this kind of impact. As much as I like Mathematica, I can't say that it shines on this front right now. And the thing that I find worrying is not that it loses the competition to C# etc, but that it loses (according to Tiobe) to Matlab and R. – Leonid Shifrin Nov 23 '13 at 13:47
  • 1
    @LeonidShifrin tks for transform in fluent English my thoughts. These are the point thats bother me. I would like to see Mathematica shines in the future, and these questions has to be overcome. All those present here have invested heavily in Mathematica, and nobody wants to contemplate the chance of him eventually become a dead or abandoned language in the near future. – Murta Nov 23 '13 at 17:18
  • 11
    @LeonidShifrin I agree with your final point. In Australia for example Mathematica gets smashed by Matlab. That in itself is not amazing because globally Matlab has had about 20 times Mathematica market share for over a decade. More "worrying" is that it also gets smashed by Mathcad, R and SPSS. The biggest problem for Mathematica has been the complete failure to develop usage in the commercial space (not withstanding some isolated examples of corporate usage). We see the results of this in recruiting where employers seek e.g. Matlab, R or SPSS skills but never Mathematica. And so on ... – Mike Honeychurch Nov 23 '13 at 23:09
14

The more I think about this, the more I agree with the "New Kind of Marketing" assertion. I also agree with other assertions that Wolfram Language really isn't a good name for what amounts to the Mathematica kernel + a few new features that we'll no doubt see in v10.

As an engineer in an industry basically run off of Excel, Mathematica really gives an upper hand in manipulating large amounts of data, and using the extensive, high-level language to present that data in forms that provide a greater level of understanding or provoke "New Ways of Thinking" (trademarked, probably). It would behoove a lot of people to at least augment their Excel usage with a little dose of Map, Apply, and f[x]:=f[x]=expr.

But in reality, I'm the only one that religiously uses Mathematica in the entire company. Not because of lack of licenses or accessibility, but because--I think--of the steep learning curve and the relative obscurity ("I didn't know we have MATLAB license here???..."). Despite the acknowledgment of Mathematica's power, there still seems to be hesitation at taking the plunge to become somewhat fluent in the language.

So, if we take Mr. Wolfram at his word, and despite the rather verbose new marketing nomenclature, if Wolfram Language increases the use of what we think of the traditional Mathematica language, then I'm all for it. Even if that means expanding the scope of questions fielded here at M.SE.

  • 2
    1+ Your last paragraph express exactly what I think. – Murta Nov 23 '13 at 2:46
  • 4
    Mathematica was probably too much a limiting name in that sense that people never took it seriously as a programming language. Any non-mathematics-related name is better. I think that both the rebranding as well as the Raspberry Pi move are great ideas to widen the audience. School science is good. Especially in poorer countries WL will be very popular within a few years. I really wished though that some useful IDE like Workbench or the Intellij-plugin would be part of WL. Or maybe should we just program by talking to Siri and WL interpetes it and generates packages and defines $PreRead etc.? – Rolf Mertig Nov 24 '13 at 1:07
  • I see value in formalising the language that corresponds to Mathematica, in isolation of any effect it may have on public opinion. One aspect is that it makes for nicer function names. Why are some functions "verby" like Extract and Take, while others like Part and Position are not? It could also really help search queries on this site if even things that are not officially implemented yet do have a name. – Jacob Akkerboom Dec 2 '13 at 14:40
6

First of all we can enjoy with expanding Mathematica territory. So far we have been ignoring questions about e.g. WolframAlpha which were not concerning Mathematica explicily. I think this policy should be still valid not only with respect to WolframAlpha but to every new technology which is somehow related to Mathematica, e.g. if one asks about the Association function such a question should be concerned to be off-topic untill it is a built-in function in next versions of Mathematica.
I guess that WolframLanguage and Mathematica overlap in 99.5%, so rather there is no real problem. Renaming Mathematica is concerned by M.SE users as a marketing trick. These comments: 1, 2, 3 by Andreas Lauschke express what do really think Mathematica users.

  • 8
    The Association function from the new doc pages looks like it's (going to be) a pretty standard part of Mathematica. If it is available right now in the RPi version of Mathematica, but not in v9.0.1, why not allow questions about it? If it's not available in any public version, then no. But if it's available in the free RPi version, then it shouldn't be disallowed on M.SE. From what I could gather online (I don't have an Pi), this "Wolfram Language" on the Pi is not really a language, just the kernel, and "Mathematica" is the kernel + Front End. I might be wrong though. – Szabolcs Nov 22 '13 at 15:42

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .