[plt-scheme] Reexpansion of modules
Matthias Felleisen skrev:
>
> On Aug 21, 2006, at 4:08 PM, Jens Axel Søgaard wrote:
>
>> Matthias Felleisen skrev:
>>> 1. I considered the idea of developing larceny modules inside of DrS
>>> back in 2002. Then I learned more about Larceny and how small it is.
>>> And we didn't have Eli's FFI yet for integrating Lareceny-compiled
>>> modules back into PLT Scheme. Now this idea is worth studying because
>>> you don't want to use Larceny for anything else than ASM. But perhaps
>>> at that level it has value! (I'll talk to Eli next week on this.)
>>> 2. Yes, we could expand PLT Scheme to PLT Scheme [core] first but
>>> even in this core language you have so many library calls and
>>> extensions, resolution of semantic issues, etc, that NOW WATCH
>>> -- compiling the rest in Larceny is either impossible or
>>
>> Is it with-continuation-mark you are thinking of?
>
> Actually Ryan moved that one into Larceny. But then there are regexps,
> custodians, wills (not lower-case W), executioners, inspectors,
> lieutenants, captains, majors, generals, and a few more friends.
I'd be content with much less. Plain R5RS would be fine for, say,
number and string calculations, e.g. an md5 computation. Well, maybe
throw in some records. I still imagine the majority of the program
to be normal MzScheme, and only write a few time critical functions
in Larceny.
>>> -- it doesn't produce code that is faster than PLT Scheme and
>>> faithful to its semantics.
>>
>> In a Larceny-as-ASM world it makes sense to tolerate a slightly
>> different semantics to gain speed.
>
> Are you willing to get #t instead of #f? 5 instead of 42? 'hello instead
> of "world"? -- Matthias
Well...
--
Jens Axel Søgaard