Ralf Hemmecke wrote:
>
> You did not answer this question.
>
> On 03/23/2014 02:17 AM, Ralf Hemmecke wrote:
> > Other question. What do you think about my introduction of
> > LispExpression. Would that kind of decoupling have a chance to go into
> > FriCAS or do you rather want to stop me thinking in this direction?
>
> I understand, that at the border between SPAD and Lisp one needs some
> types that map one-to-one and the compiler and runtime system should
> know about this. So I fear my idea of decoupling by introducing
> LispExpression is not helpful in any way.
>
> It's just that it looked better to me if there were no
> interdependencies. Unfortunately SPAD has no "extend" so that one could
> first introduce a data structure
>
> Integer: with == add
>
> maybe with only a few functions, but all only depending on the
> underlying runtime system (Lisp or whatever), but not on any other SPAD
> type and then "extend"ing later to the respective category.
I somewhat got used to idea that there are recursive dependencies
between types. So I do not feel much need to break things into
layers to eliminate recursion. OTOH I took a look at Aldor
libraries, and IMO 'extend' make things less clear. With
'extend' writing code is easier, especially it gives power
to users to get new functionality without touching established
libraries. But functionality is spread into multiple
extensions and that makes reader task harder.
So I do not like the idea of spliting basic types. We probably
will be forced to do some splitting, to simplify compiler
bootstrap, but I would like to limit splits to necessary cases.
--
Waldek Hebisch
[email protected]
--
You received this message because you are subscribed to the Google Groups
"FriCAS - computer algebra system" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/fricas-devel.
For more options, visit https://groups.google.com/d/optout.
