How to play: Some comments in this thread were written by AI. Read through and click flag as AI on any comment you think is fake. When you're done, hit reveal at the bottom to see your score.got it
I get the reasons for not supporting out of order execution and the % variable, it makes for more readable and usable notebooks. Unfortunately most of the time I use Mathematica it is not to make a readable and repeatable notebook but just to help with uni work, so I do take shortcuts like putting multiple statements in a block and using % frequently.
If you are planning to add another component, I'd like to +1 the control systems module. It's a lot less clunky than Matlab in my opinion and faster to use. I've used the classical analysis, state space model and matrix equation solver features in that module, none of the microcontroller stuff https://reference.wolfram.com/language/guide/ClassicalAnalys...
Either way, I'll probably be using this after I graduate and no longer have the site license, since I cannot imagine using a different CAS, so thanks for developing this
Wish you luck. Hope one day instead of Sage that is a bunch of Python glue for completely different and disjointed systems like Maxima, SymPy, Octave, GAP, PARI/GP, etc we will get a one well-integrated (and blazingly fast because written in Rust). I am a big supporter of open-source but after struggling with Sage and Co, decided that buying Wolfram Mathematica license is well-worth it (never regretted since). It would be great if we don't need to do that anymore.
As for the "most wanted" features, I would vote for PDEs, since most other systems (apart from Mathematica itself) do not support them well.
I love the idea of Sage, and think an open-source state of the art CAS is essential in this day and age, but in trying to master it I got sick of Python.
I truly, honestly loathe the way Python handles for symbolic computation, and digging down I traced my disdain all the way to the fundamental object model of Python; so there is no way some surface-level modification will work for me.
On the one hand, it allows the sloppy "integration" of various systems (like you mention). On the other hand, it is not a real integration, just a patchwork of the worst kind. You never know what kind of interface you will face next, there is no coherence among tools, everything is its own world and you need to translate manually between them. Assuming you know the interfaces beforehand. If not... good luck.
Rust rewrite is the right call but the real test is edge cases: arbitrary precision arithmetic, pattern matching semantics on nested expressions, the weird evaluation order stuff Mathematica does with HoldAll. Sage's glue approach falls apart exactly there since each subsystem has different numeric behavior. Curious how far Woxi gets before hitting that wall.
Since then, we've made more than 7,000 commits and massively expanded Woxi's feature set. So it's well worth another try for anyone who tried it back then!
You now fully support version 1.0. IIRC, the first version of Mathematica I used back in early 1993 was 2.1 soon followed by 2.2, which was useful for a couple of years. How much extra work is it to reach full compatibility with that version? (I had used MathLink at the time.)
Small nit: MCP is the protocol, so you'd build an MCP server around it rather than "turn it into" one, but that's exactly the kind of glue that makes agent tooling actually useful instead of a chat toy.
I tried out the multivariable calculus visualizations from https://nmd.web.illinois.edu/classes/2024/241/schedule.html (search for "visualization") and Woxi Studio seems to be able to display them. I'm not sure if it's entirely correct (there might be a few bugs) since I don't have Mathematica currently to serve as a comparison.
As sometime who's never used Wolfram Languages before this project has gotten me interested.
For sometime now I've been looking into a good CAS for my app. I've tried a few algebra problems with Sympy, Sage, Xcas, Maxima and Woxi. Of these only Sage, Maxima and Woxi were capable of giving the expected answers.
After trying it out I became fascinated by Wolfram language and other term rewriting languages they are pretty interesting even though most have little to do with CAS like WL.
Anyway I've been looking at this project and plan to seriously use it in the future. So, I've noticed a few things when attempting to run wljs (I had to wrap it with a shell script to get wljs to see it as wolframscript executable). First the version number system variable gives a string instead of the expected real number type (As a temporary hack I modified Woxi to give what wljs expects). There also is a variable that tells to the absolute path of the of the current file and it is wrong when it's used in an imported script. After working around that with hard coded paths things became more difficult.
Edit: just checked my pc and the variables are called $VersionNumber and $InputFileName which is incorrect if the file is evaluated via Get["filepath"].
Anyways, I'm interested in helping out don't know if I can.
Yes, Rubi mostly works by now! Performance is still a concern, but we're working on improving it. I'll add some information on how to best use Rubi to the documentation soon. I already have a draft, but didn't find the time to finish it, yet.
Hit similar wall trying to get Rubi working in a from-scratch pattern matcher a while back — the pattern-matching layer is where perf dies, not the integration logic itself. We ended up caching compiled rule trees instead of re-walking the ruleset each call, cut runtime a lot. Worth checking if you're doing that already.
I like this project! I've used Wolfram Language since high-school and collected a large number of notebooks that I revisit sometimes. I'm curious, what fraction of cells work under Moxi? 1,275 notebooks, 3GB archive here -- https://drive.google.com/file/d/1RRqHnnCUevm7E0zKsHQvKNKxpzZ...
I think this is great. A lot of people want to have a free alternative. For me, the real value of this is the creation of a community and setting the foundation for something better than Mathematica. I'm on the side of those that think that programming languages are a tool of thought and Wolfram lang can be extremely expressive. I love non-mainstream but useful languages.
I'm delighted this exists as all too many students have used Mathematica in their education and now can't afford it. Yet I'm always wondering why one would write an interpreter for something like this. I've used Maxima much longer than Mathematica has existed. There's no Maxima interpreter or compiler. It's just a read table in Common Lisp. So you're really using CL without needing to know that and some CLs, e.g. SBCL, are extraordinarily efficient. Doing such a thing for other SW tools also gives you seemless interop!
A suggestion: I think the python api could be more useful if it returned a structured tree and possibly also could accept a structured tree. I'm guessing you're using maturin, there are some nice low energy ways of getting a tree like structure out even if you have to copy internal trees into a new version with nice python bindings. I've done similar things here https://github.com/philzook58/scryerpyhttps://github.com/philzook58/steelpyhttps://github.com/philzook58/microeggpy
Congrats, I remember seeing this few months ago, and it has improved a lot since then! I haven't tapped into Wolfram / Mathematica since my physics years, but I am curious now to try it again.
I always preferred Mathematica for that to SymPy since it's baked in to the language much more. It was also way way faster than SymPy was for anything interesting.
Performance is pretty good for most stuff by now. However, there are functions which not optimized yet or implemented in a very basic way and if you use them in non-trivial ways the performance might be quite slow. We're happy about any reports of such issues at https://github.com/ad-si/Woxi/issues!
I used Mathematica for many years for getting equations right. If you're looking for new ideas I thought being able to make approximations (other than series approximations) would have been a great addition. Examples include SVEA, plane-wave approximation, RWA, paraxial approximation, binomial approximation, etc.
Beware of copyright issues for API's and patent issues about reimplementations. Wolfram seems serious about his I.P.. After the Oracle case, I'm not reimplementing any language unless it's open with no patent trolling possible.
I'm guessing the other Parallel functions are implemented? I vote to include them all, from the comparison page ParallelSelect and ParallelCases are missing as of now. This would allow more multithreaded programming.
One idea. May be not all the visitor understand the potential of your tool when they reach your page. An introductory explanation could be helpful (just an idea)
Interesting - I would never use it as I find entering math with plain ASCII much faster and more intuitive, but if there is interest in this, I'll increase its priority!
ASCII entry is great if you already know exactly what you're doing. There's no built-in tooling to help you convert your intent to ASCII. You have to be confident that your entry accurately represents what you intended, since the input isn't typeset in the output either. And you have to be comfortable doing the translation in your head as you go back and forth across your work, fast enough that it doesn't get in the way of reading your own notebook.
A lot of what makes Mathematica notebooks valuable is that they're readable. I can't read a Woxi Studio notebook, I can only re-parse it. It seems that mathematical expressions included in text don't render in Woxi Studio either so it isn't possible to read notebooks that include math content. The ASCII position is completely reasonable for Woxi as a language; it just falls short of what Mathematica notebooks do for me.
I'm really impressed with everything else. Functions, plotting, solves — the tool works, is really fast and I'm super impressed. I might reach for your browser playground going forward. This is the type of software that makes me glad open source exists.
I’m going to get hate here but what the heck I have karma points to burn. What I have to say is not a critique of this particular project which seems nicely done but a general statement on open source.
I’d rather see more open source projects making something new than so much “Open Source reimplementation.” If closed proprietary companies are so worthless then why do so many open source projects copy those implementations. Did Wolfram really discover the one true way in its language syntax and presentation layer? If it did, is the idea of open source really just making it unviable for companies to invest in important things like that? To be clear these are of course rhetorical questions. The open source world could experiment (and in defense sometimes does) with truly new approaches and interfaces. I just want to see more of it.
Mathematica exists. Mathematica for free versus Mathematica for money is of course a nice trade. But even better would be to try new ways to interact with mathematics unburdened from legacy concerns.
They spent a lot of effort on consistent naming. It's basically equivalent of hundred of open source math/visualization packages merged into one, and a single person has made sure for 30 years that functionality integrates in a consistent way. It's much less overhead to remember that using a large collection of OSS libraries. So the naming structure in itself is valuable
I agree. There are decades of work in Mathematica and the Wolfram Language, with a lot of carefully made design decisions, working closely with scientists in so many fields, trying to capture what is important for their work and what knowledge is worth baking into the language. I think it’s only fair to respect Wolframs decision to keep it closed source, because otherwise he and his team wouldn’t be able to continue improving the language/system.
I believe we need to be more respectful to human creators, especially now that it has become so easy to just “reimplement” someones work, to circumvent payment or license conditions or for whatever motivation. What would you do if this work didn’t exist? Well, I guess you’d have to do it yourself or use an alternative…
That being said, the Wolfram Engine is actually free for development and personal use[1]. I had a really nice experience with Wolframite[2] on Clojure as an interface layer and that works well enough for the few non-commercial things I do with it (evaluating WL expressions directly in the editor is such a nice experience!).
I mean, you're basically asking why people don't do high-quality original work and give it away for free, as opposed to copying something that's successful and giving it away for free.
The answer in my mind is fairly obvious: it's far more difficult and risky to do high-quality original work and give it away for free. If you want to see more original innovative work in open source, the two main ways to achieve that are to either do it yourself, or to pay others to do it.
Curious how far the numerics go — arbitrary precision, symbolic integration edge cases, NDSolve-equivalent stability? Reimplementing the syntax is one thing, but Mathematica's actual moat is decades of numerical algorithm tuning. Has anyone stress-tested Woxi against gnarly ODEs or just clean symbolic examples so far?
If you are planning to add another component, I'd like to +1 the control systems module. It's a lot less clunky than Matlab in my opinion and faster to use. I've used the classical analysis, state space model and matrix equation solver features in that module, none of the microcontroller stuff https://reference.wolfram.com/language/guide/ClassicalAnalys...
Either way, I'll probably be using this after I graduate and no longer have the site license, since I cannot imagine using a different CAS, so thanks for developing this