Glen Whitney

Best-laid plans and all that -- that was over a year ago. Can't say I know when/whether I'd get back to this, but it would still be a good change.

As #2290 has been merged, I believe this is completed. Closing.

This is a very valuable idea, and has spurred multiple POCs (including one I did myself), but still is rather more of a Discussion than an issue. Hence, moving.

Looking at the literature, one simple choice (that used to be commonly used, but apparently has been superseded in more recent even better algorithms) is just to take k to...

OK will do that after I respond to all the code review in #3037. Yes we still need precision, convergence in eigenvalue problems is always touchy. My understanding is that...

Oh and on precision: it's also totally necessary with bignumbers, and I think always will be.

Since the flag to only find eigenvalues is already there in `src/function/matrix/eigs/complexEigs.js`, it would only take a small change to the definition of the `eigs()` function to enable this. (The...

Sorry, the proposal here is unclear to me: Isn't the question of whether an Error string can contain Unicode a JavaScript question, not particularly a mathjs question? Or is there...

OK. As you probably know, in the _specific_ case of undefined symbol and undefined function errors, you can simply redefine `math.SymbolNode.onUndefinedSymbol` or `math.SymbolNode.onUndefinedFunction` on the fly, e.g. ``` math.SymbolNode.onUndefinedSymbol =...

This is definitely a bug, and it definitely still exists. In fact, it's worse, in that if M is a matrix, `null > M` **also** reports `TypeError: Unexpected type of...