newRPL: DATA TYPES USING [] AS ELEMENT CONTAINER

[Claudio L.]

(04-14-2017 03:23 AM)compsystems Wrote:  I hope your comments, problems of interpretation by the compiler, inconsistencies, ambiguities, other types of data that can be included etc.

They like the idea?

Thanks

I think you misunderstood the problem and took it too far in your own direction. The problem is not the choice of a particular syntax, it's that allowing those objects in a symbolic expression turns the expression into something else, with a huge number of special cases that the CAS will have to (magically?) handle.

Your FACTOR example: the factor command would have to work with symbolic expressions and with those polynomials in vector form embedded in the symbolic expression as a special case, and then the mixing of thos vectors and symbolic terms, and then also any other strange object that you embed into the symbolic as well. Calling it p[] or pl[] or v[] doesn't solve the main problem at all. Still, every command and the CAS will have to deal with all those objects independently, intermixed, etc. It's a coding nightmare. The idea is that unless we can figure out a way for the CAS to treat those cases uniformly there's no point in choosing a particular syntax.
Your syntax is not bad but doesn't add anything that could help solve the problem, and I also see it as a departure from the RPL syntax we all know. Your examples also do a lot of numeric work using algebraic expressions: that's what you'd do on a Prime, not on an RPL machine.