I think that (part of) the point is to be able to bootstrap with minimal tool support, and thereafter to be self-supporting. The bootstrap process is described better than I can in Section 6.1 of "Accessible Language-Based Environments of Recursive Theories", available on the VPRI website. C is described as a "portable high-level assembly language".
Cheers, Josh Kjell Godo wrote: > Wouldn't this whole thing be better written in Smalltalk than in gcc? > > gcc is more portable? > > Couldn't Smalltalk output assembler that gcc could then compile? > I saw an assembler written in Smalltalk once. Hope I still have it. > It was very simple and clean looking. > > I am writting a compiler for a language called picoLARC and I am > doing the language specification in Smalltalk which compiles it to > trees of objects which are then run/evaluated by a recursive method. > I'm getting the ideas straight in Smalltalk first. > > Then I suppose I would rewrite the picoLARC system in picoLARC > on top of Smalltalk. It would be slow running but what do I care about > that? The picoLARC compiler etc gets translated into executable trees > which then translate themselves into assembler or machine language or > IL or whatever. > > These executable objects aught to be able to generate linearized > code like assembler or machine language. According to the book > Scheme in Small Pieces. > > Perhaps the executable objects could generate a C# program that > was nothing but op code generation statements and then picoLARC > could run on and access .net for windowing and stuff. Or it could > run in the browser like Vista Smalltalk( I think it's called Vista ). > > Wouldn't it be a lot easier to do fonc in Smalltalk? That way you > could map out all the VTables and everything in a quick way that > would be easy to follow. Then this Smalltalk thing could generate > assembler or whatever which could be compiled into a something > like running Smalltalk but when you open the browsers all the code > is cola jolt fonc or whatever you call it. And that thing could then > generate headless( no GUI ) executables if needed. > > Smalltalk is reflective. gcc is not. So what is the argument in favor > of gcc? Why was it chosen? Is programming in gcc faster than > programming in Smalltalk? > Easier to understand? Better documented? Self documented? > Easier to debug? > > If I could implement picoLARC on top of COLA or whatever it's called > that would be cool. But the idea of trying to program in gcc always > stops me dead. If cola or a version of cola was like Smalltalk but > removing the problems that Smalltalk has then that would be good. > > If I could make the picoLARC compiler in Cola and build up a picoLARC > that was like Smalltalk but was fast where it needed to be fast and > could access .net or the various libraries in Linux that would be good. > > I would like to request a version of Cola that can be used to make a > new language that is as fast as one made with gcc where it needs > to be fast and can access all the libraries. And that Cola would > have a Squeak like GUI that was detachable with all the browsers > in it. And it could generate all kinds of different executables in easy > ways. It could generate a VM like Squeak or a lib with no GUI. > > On 6/22/08, Ian Piumarta <[EMAIL PROTECTED]> wrote: > >> On Jun 22, 2008, at 2:41 AM, John Leuner wrote: >> >> >>> What do you use a teletype for? >>> >> Backing up my sources to punched tape. >> >> Cheers, >> Ian >> >> PS: if you're taking me seriously in this tread, it's time to stop. >> >> >> _______________________________________________ >> fonc mailing list >> [email protected] >> http://vpri.org/mailman/listinfo/fonc >> >> > > _______________________________________________ > fonc mailing list > [email protected] > http://vpri.org/mailman/listinfo/fonc >
_______________________________________________ fonc mailing list [email protected] http://vpri.org/mailman/listinfo/fonc
