newRPL - build 1255 released! [updated to 1299]

[Claudio L.]

(12-19-2018 08:04 PM)The Shadow Wrote:  Also... Do we actually *need* to be able to handle matrix equations? OldRPL can't.

Good point, perhaps we don't. However, if we are designing a new CAS from scratch, I'd like to make it useful not just for a kid to get a result that looks like a book, but for more advanced users to make it their own as well.
You mentioned you liked to use the CAS to simplify boolean operations, that's an actual real application, and I'd like this CAS to be able to do it without anything special. Just define a few rules and you should be good to go.
Back in the days of the hp48 I did a lot of work with matrix expressions, and I could've used a CAS to simplify some proofs.
The main thing is that I don't want the automatic simplification to mess up the equations (like putting a 1 where it should've been I, or doing permutations on multiplicands that could be matrices), to at least give a chance to those future user-defined rules to work.

You threw an idea of using names like X.MAT. That's basically what the Prime does, variable names that imply the content based on the letter. What I don't like is that forces one to use an unnatural selection of names on equations. That maybe OK for doing homework, but in physics you want to call variables by their "usual" name (P*V=n*R*T wouldn't mean the same if you change the letters, would it?)

I think we need a way to display the attributes of variables, though. I'd like a variable to be called the same, but look different based on its attributes. For example, bold text might be constant variables, perhaps complex ones could be double-struck, and matrices and vectors could have a line over them (like complex infinity does). We'd need a lot of work on fonts though.