[Uim] Proposal for SigScheme

TOKUNAGA Hiroyuki tkng at xem.jp
Fri Jul 22 02:53:13 EEST 2005


On Fri, 22 Jul 2005 08:20:16 +0900
Kazuki Ohta <mover at hct.zaq.ne.jp> wrote:

> Hi, All.
> 
> I receive the proposal for SigScheme as follows. This is 
> actually the contents of branches/r5rs/sigscheme/MEMO file.
> 
> ====================================================
> > These issues should be resolved until 0.5.0. I'll remove 
> > this meme until 0.5.0 release.
> when do you plan to release 0.5.0?
> 
> > * require/load
> > In gauche, require works at compile time, not at 
> > executiontime. SigScheme should be so. Instead of require,
> > load actually works at execution time. 
> > Gauche uses dynamic-load to load shared library. 
> OK. I confirmed. The last line mean that you want 
> "dynamic-load"?

Maybe. I don't know for now. It's only memo.

> # I don't know "require" procedure's definition, it's not 
> # defined in R5RS. Do you know?

No, I don't know too. But SLIB also implementing require macro. Maybe
it be another reference.

> > * vector
> > Vector will be very important data structure. Sorted 
> > vector data will make uim faster. In addition, it would 
> > reduce memory consumption. 
> > In Gauche, find, map, for-each, and remove are applyicable 
> > to vector. This feature need generic function,  I think 
> > generic function is not  need for SigScheme. So I'll 
> > propose 'vector-' prefix to impelement map, for-each, 
> > remove for vector. i.e. vector-map is map for vector. 
> > * define-record-type
> >Defined as SRFI-9, define-record-tipe should be implemented 
> >to replace current define-record.
> 
> This proposal relates to SRFI. I'm now considering about 
> whether SigScheme implement it. I think SigScheme's core 
> should be small. So, I plan to implement SRFI's procedures 
> in Scheme code as much as I can.
> 
> How do you think about this?

Sometimes Scheme code consumes much more memory than C, so 
that depends on the situation.

In addition, development cost is also important. If writing in Scheme
is considerably easy, Scheme code sould be used. Rewriting with C later
is always possible.


Regards,

-- 
TOKUNAGA Hiroyuki
tkng at xem.jp



More information about the uim mailing list