Re: [Axiom-developer] Re: SPAD and Aldor again

[Gabriel Dos Reis]

C Y <address@hidden> writes:

| Gabriel Dos Reis wrote:
| > C Y <address@hidden> writes:
| >
| > | Indeed.  The Aldor documentation is not free at all, and any attempt to
| > | define Aldor in a literate style would have to duplicate Aldor without
| > | duplicating too closely its documentation - that's a real problem.
| > 
| > I cannot parse this. Could you elaborate?
| 
| Sorry, this seems to be an off week for me communication wise.  If we
| define and document properly a language like Aldor, we must include
| descriptions of the behavior of the language.  The problem with
| specifying a language behavior is that the description cannot vary too
| much from one description to another and still be defining the same
| concepts.  The problem of (for example) writing a document that
| describes ANSI lisp without running into any copyright problems is very
| difficult, because if you stray too far from the text of the
| specification you run the risk of not defining the behavior you need to
| define.  And it would be even more difficult to avoid such a description
| being a "derivative work" of the previous description.

OK, thanks for the explanation.

Since I'm not in the business of cloning Aldor, I'm not sure how that
affects Axiom.
I don't see a point of cloning Aldor.  
I see great benefits in an improved SPAD.

[...]

| > You don't need a language to do *just* mathematics in Axiom.  You also
| > need a language to communicate with the world around.  All major
| > systems for computation mathematics have grown into that position.
| > Don't get blindsighted.
| 
| I would prefer to let the Lisp level handle the outside world as much as
| possible.

But you still need to specify what that does to a SPAD program.  Just
delegating to Lisp does not solve the fundamental problem.

|  What are you referring to by communication?  Exporting
| algorithms as Fortran code?  Interacting with C libraries?  File and
| Data Input/Output APIs?

All of that, including interfacing with nay reasonable language used
in the computational science community -- that list goes beyond
Fortran and C.

-- Gaby