Re: records

[Michael Livshin <mike@olan.com>]

ttn writes:
> so my question basically is: what is the roadmap for guile records?
> specifically, what does the future hold for the following procedures:
> 
>         make-record-type
>         record-constructor
>         record-predicate
>         record-accessor
>         record-modifier

you, of all people, should know about the aptness of the word
"roadmap" when talking about Guile.  ahem ;).

now, I don't think you should be afraid about the future of the above
procedures.  even though SRFI-9 describes only the declarative syntax
for record type definitions, I'm sure that any real-world
implementation would provide the procedural interface as well.  you
might have noticed that the SRFI-9's reference implementation uses the
exact procedures you are worried about (and, in fact, defines them
first, in terms of vectors).

as for the ways of implementing records for Guile, well, you could do
it in Goops, or you could do it with structs (as it is now).  I have
some great plans to replace structs with suitably extended version of
records, and even have most of the design in place (if my mind can be
called a place ;), but I ever lack time.

so to present the things in some order: SRFI-9 API can (and, IMHO,
should) be implemented with the procedures you ask about.  these
procedures are there now, and will remain as long as structs remain.
and if I finally do something about structs, the procedures will be
reimplemented compatibly.

NB: the above is merely my understanding of the situation, I don't
speak for the guys who happen to actually write code, etc.

--mike