On Mon, Sep 18, 2017 at 6:04 PM, Ralf Hemmecke <r...@hemmecke.org> wrote:
>> Since data in Lisp is already tagged, we can always prevent
>> unnecessary memory allocation, for Integer, we can use NIL
>> to indicate failure,
>
> Really? The (lisp) integer 0 and NIL are distinguishable?
> Does that work for all implementations of lisp? I would bet that this is
> an implementation detail of the specific lisp fork (sbcl, ccl, gcl, etc.)
NIL is a symbol, it's always different from integer 0 in Common Lisp.
> There's only one big BUT. I don't like so much to make the FriCAS
> algebra dependent on LISP details. IF this is done, then it must be done
> in a very restricted area of the code. In general the Algebra code
> should have zero knowledge about whether it runs on top of C, Lisp, or Java.
>
> Ralf
First we can always have Rep := Union("failed", R) as default
implementation (which is generic), and second, I believe for
other target languages, there are similar tricks.
For Lisp, since it's already a dynamic language and has tagged
data type, we should take advantage of that and should not tag
data again (by Union("failed" ...)).
--
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 fricas-devel+unsubscr...@googlegroups.com.
To post to this group, send email to fricas-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/fricas-devel.
For more options, visit https://groups.google.com/d/optout.
