Re: [Axiom-developer] Curiosities with Axiom mathematical structures

[Gabriel Dos Reis]

Martin Rubey <address@hidden> writes:

| Dear Gaby,
| 
| I just would like to say that I do follow your reasoning and I do
| think that we should consider whether it makes sense to adapt Aldor
| to cope with the situation under discussion.

Many thanks!

| However, I do believe that this would imply a lot of work, and we
| really should see how often this situation occurs in practice. Of
| course, it is sad that it does ocur already so deep down in the
| hierarchy, but if this is the only place where we are in trouble,
| maybe we should ignore it for the moment. 

I agree.  It is a hard place to leave with history -- that would be my
second summary of my involvement in evolving languages like C++.  
Nevertheless, if we cannot attain the ideals, we should try hard to
approximate them to the best of our capabilities.

| Do you have experiences with hacking compilers and the like?

yes, I have been working on GCC (mostly the C++ front-end and library)
for nearly a decade now.  I also have some experience in
program transformations.

| Do you think you could implement the necessary changes?

I don't know the internals of the Aldor compiler, so it is hard for me
to say how much of work it would take.  I suspect 'a lot' especially
since we do not have a "formal" definition of the language.

| Did you read the proposed solution by Nicolas Doye?

I have not get a chance to read his thesis -- however, I'm printing it
while I'm composing this message.  I'll get back to you.  
The next week and the one after will totally filled up for me by
daytime job and work on getting "concepts" for C++.  However, I expect
to send feedback relatively soon (assuming planes are delayed :-))

-- Gaby