williamstein

What is the DeepSeek team? Who is making this?

Coincidentally, I'm in the middle [1] of building something for https://CoCalc.com that is exactly what you're describing. For collaborative document editing (e.g., google drive and overleaf) it's a common feature, but for code editors it isn't. CoCalc is both. Anyway, nothing to see yet, but you might want to check with us in a month. After thinking about this problem a lot recently, I think it’s critical to store the comment locations with all versions of the file, so you don’t lose comment locations, or at least maximize the information you have available to locate comments when they get lost.

[1] https://github.com/sagemathinc/cocalc/pull/8071

Kevin Buzzard

It is a thing. Your experience doesn’t at all match my experience driving on the West Coast for the last few decades.

> “Our main product was an SDK Generator that could take any OpenAPI specification and generate high-quality client libraries in multiple programming languages.”

This is something I could really use! Is there a product that does this and doesn’t cost an arm and a leg?

And the professor who taught that course won a major prize today: Ribet to Receive 2025 AMS Steele Prize for Seminal Research - https://www.ams.org/news?news_id=7391&fbclid=IwZXh0bgNhZW0CM...

I took that class with you! It was amazing. Here are my notes: https://wstein.org/books/ribet-stein/

If the BSD rank conjecture were false, then the simplest counterexample might be an elliptic curve with algebraic rank 4 and analytic rank 2. This could be established for a specific curve by rigorously numerically computing the second derivative of the L-series at 1 to some number of digits and getting something nonzero (which is possible because elliptic curves are modular - see work of Dikchitser). This is a straightforward thing to do computations about and there are large tables of rank 4 curves. This is also exactly the problem I suggested to the OP in grad school. :-)

In number theory doing these sorts of “obvious computational investigations” is well worth doing and led to many of the papers I have written. I remember doing one in grad school and being shocked when we found a really interesting example in minutes, which led to a paper.

This is really amazing as is their standalone pglite project, which is a WASM port of Postgres. One surprise for me though is that Electric is a read only replica. In particular, “Electric does read-path sync. It syncs data out-of Postgres, into local apps and services. Electric does not do write-path sync. It doesn't provide (or prescribe) a built-in solution for getting data back into Postgres from local apps and services.” https://electric-sql.com/docs/guides/writes

Pure Python is on the order of 1000x faster at computing that sum by brute force than Mathics. This suggests that perhaps some of the basic optimizations one does when implementing a language, to get even the most minimal level of performance, haven't yet happened with Mathics. For example, when we added asymptotically fast arbitrary precision integers to SageMath (by wrapping GMP, then later MPIR), we had to implement an "integer pool" since otherwise a lot of everyday microbenchmarks were far too slow.

Note: Mathics is an optional Sage package - https://doc.sagemath.org/html/en/reference/spkg/mathics.html...

Wow, they do public issue tracking in an unusual Way, via a mailing list: https://www.postgresql.org/list/pgsql-bugs/

And to read the source code.

When you go into battle to solve a computational mathematics problem, the problem sometimes doesn't care about these boundaries. E.g., you might think that you're solving a problem in "computer algebra", but a sufficiently fast solution might end up involving numerical linear algebra in surprising ways. (E.g., enumerating elliptic curves over the rational numbers is reduced to exact linear algebra because of a Wiles work on Fermat's last theorem, and often exact linear algebra can be done via clever tricks much more efficiently using floating point matrices, which happen to be insanely fast to work with due to GPU's...). It's thus valuable, at least for research, to have large mathematical software systems that span these boundaries.

I'm not aware of any real push to include Mathics in sagemath. I speculate it's because SageMath is mainly developed by research mathematicians and cryptographers, so performance is often a primary concern when evaluating components for inclusion. One of the reasons SageMath is the largest Cython project is because Cython makes it possible to utilize fast C/C++ libraries from Sage.

I have the impression that Mathics is currently not seriously concerned with performance. E.g., try this little microbenchmark in Mathics: "AbsoluteTiming[Sum[i, {i, 1, 100000}]]" or read their roadmap. This is fine -- there are many interesting applications of the Mathematica programming language where performance isn't important, e.g., carefully stepping through some symbolic operations with expressions. However, for Sage the motivation of most developers is cutting edge research math, and for that, performance is almost always very important. Performance is why Sage also implements a lot of similar functionality to Sympy rather than just using sympy for that functionality -- since sympy can be relatively slow due to their priority to be easy to install, which is definitely not a priority for SageMath.

The mission statement for SageMath is to be a viable alternative to Mathematica, Matlab, Magma, and Maple, but I never meant that to mean being a clone (e.g., directly running Mathematica code), but instead just an alternative in the sense that it can support research built on open source mathematical software that might otherwise be done with those closed source programs.

Proving that there exists an algorithm that results in an eventually consistent view of the document is trivial. However, that's not what we're talking about. Instead, researchers define a specific algorithm (e.g., involving CRDT's or OT's or something else), then prove that their algorithm results in an eventually consistent state. This reminds me a little of the relationship between proving that there exists an algorithm to factor all positive integers (this is trivial) and proving there exists a subexponential time algorithm to factor integers, which is much less trivial (see https://en.wikipedia.org/wiki/Lenstra_elliptic-curve_factori...).

What they claim is that if all editing stops, then after a period of time everybody will be looking at the SAME document. This is what is meant by "eventual consistency". Achieving this in general is indeed a difficult problem, but (some of) these algorithms do solve it, though it can be tricky to correctly prove that they do. I agree that it is not possible to ensure that the document everybody is looking at is what they actually wanted it to be. However, there are some options for what happens, where some results may be technically correct -- we are all looking at the same thing -- but obviously really bad. This beautiful talk has some examples: https://www.youtube.com/watch?v=x7drE24geUw

The text sync algorithm I wrote succeeds at the "The Color of Pomegranates" test. See https://cocalc.com/wstein/dev/share/files/2024-12-06.md for some details of the exact patches. This algorithm, which we came up with in 2015, is described at https://blog.cocalc.com/2018/10/11/collaborative-editing.htm... It's a different approach than that taken by Yjs, Automerge, and Egwalker, with signficant pros (and cons!) relative to those algorithms. It has been used in production in CoCalc for over 8 years, and was also used by the recently shutdown Noteable Jupyter notebook project.

“Americans born before 1996”

I switched to using Firefox instead of Chrome and now "=" works! :-) I'm using GFA Basic from https://www.atarimania.com/pgesoft.awp?version=29409

Within seconds I found GFA basic (which I spent endless time programming in as a kid), and got it running. Unfortunately, I can't figure out how type an "=" (equals) character from my Mac laptop.

Does anybody know what the analogous story is for Google Cloud Storage? It seems to me to be unlimited with no per bucket cost, but I never found a definitive statement about this. It’s important for SAAS products with one bucket per customer…

No mention of Kubernetes (and k3s, etc.)

- CoCalc (which I wrote) can be self-hosted but is NOT FOSS, though you can browse the code here: https://github.com/sagemathinc/cocalc -- do a quick "self hosted" test if you want by typing "docker run --name=cocalc-trial -d -p 4043:443 sagemathinc/cocalc-trial"

- Google Colab is fully closed source and I think there is no self hosted option.

- Deepnote is fully closed source and I think has no self hosted option.

- Saturn Cloud enterprise is a self hosted option for Saturn Cloud. Saturn Cloud is I think a nice management layer around open source JupyterLab.

- I think Apache Zeppelin is self-hostable FOSS.

This is a blog post by ClickUp -- it's basically an SEO article "10 Jupyter Notebook Alternatives and Competitors in 2024" published two days ago, which isn't badly written. CoCalc, Colab, Deepnote and Saturn cloud are all on the list, and skimming it I didn't see anything really wrong with their descriptions. The description of their own product ClickUp seems odd, making it out to be a Jupyter competitor, which seems a bit of a stretch. ClickUp appears to be a fancy collaborative AI assisted project management and work tool - it's pretty cool, but nothing like Jupyter.

This is seriously triggering my CoffeeScript PTSD.

I recently discovered Rspack, which is a compatible rewrite of Webpack in Rust by a guy at ByteDance. It is genuinely 5x-10x faster across the board on my large/complicated project. I've been using Webpack for 8 years, and I was absolutely blown away to be able to easily swap Webpack out for something so similar (written in Rust) and get such a massive performance improvement. This has made my life so much better.

Matt said in his keynote that he had a kidney stone a few weeks ago, which is evidently extremely painful. Perhaps that physical trauma triggered something.