Re: [Chicken-hackers] R6RS

[Elf]

Chicken is more alive than ever, so it's not a good comparison. :)

I personally have been working for some months to make Chicken work correctly
on basically every arch/os combination known to man, and to start abstracting
the compiler layer from the library layer (so that we can do, say, a jvm
backend, or a native objc backend, or what-have-you).

Additionally, I've been working on native threading. :)

Stalin didn't die due to being R4.  There are still plenty of R4 implementations
out there perfectly alive. (MIT, SCM, Guile (until recently), Tinyscheme
(probably the most widely used scheme), to name but a few.)
Stalin died cause it was hard to use, took forever, and because the code
it generated, while sometimes good, was impossible to do anything with.

Chicken is R5 and will stay that way until R7, at the very least.  We've 
taken a few concepts from the R6 discussion (such as a module system) and 
implemented them.  (FWIW, in my own discussions with some of the strongest R6
backers, that chicken implements a module system now means 'r6 was a success'.)
Not all the concepts are bad.  However, the underlying PHILOSOPHY of R6 is 
fundamentally wrong and flawed, and that is why R6 is not, nor will ever be,
an option for chicken.  R6 is not a set of language features; its a philosophy
of how Scheme should be, and we cannot agree with that.

-elf


On Mon, 3 Aug 2009, Thomas Bushnell BSG wrote:

> It's like I didn't even say.  Oh, I did.
>
> Stalin decided that r5rs features were unnecessary to implement, and
> dropped them.  Alas, Stalin is dead.  I don't want Chicken to die.
>
> Thomas
>
>
> On Mon, 2009-08-03 at 19:39 -0700, Elf wrote:
> > Out of curiousity, what good reasons are there for using R6, since you seem
> > to be so behind doing so?
> >
> > -elf
> >
> >
> > On Mon, 3 Aug 2009, Thomas Bushnell BSG wrote:
> >
> > > I'm sorry; those are reasons not to approve r6rs.  There are many
> > > things, it seems to me, which don't belong in the standard, but which
> > > should be in any good Scheme system.
> > >
> > > Are you saying that if X should not be in the standard, then it is wrong
> > > for Chicken Scheme to implement X?  Suppose r6rs had mandated the
> > > Chicken Scheme FFI; would you begin lobbying for its removal from
> > > Chicken Scheme?
> > >
> > > Thomas
> > >
> > >
> > > On Tue, 2009-08-04 at 08:54 +0900, Ivan Raikov wrote:
> > > > There are many good reasons not to support R6RS. You might want to read
> > > > them here:
> > > >
> > > > http://www.r6rs.org/ratification/results.html
> > > >
> > > >
> > > > Thomas Bushnell BSG <address@hidden> writes:
> > > >
> > > > > So the extraordinary compiler Stalin is sliding into obscurity, because
> > > > > it remains stuck at r4rs, with nobody having done anything to bring it
> > > > > up-to-date.
> > > > >
> > > > > Chicken Scheme can be talked into being r5rs, which is great, though
> > > > > it's not obvious sometimes what needs to be done.
> > > > >
> > > > > But r6rs has been out a while.  I think we should create a "punchlist"
> > > > > of to-do items, with the goal being to get r6rs compliance in the
> > > > > not-too-distant future.  With the version 4 macro system in place, this
> > > > > shouldn't be too far off.
> > > > >
> > > > > And yet, we have a quote from Felix Winkelmann from a couple years ago
> > > > > saying it will never happen, and "R6RS must die".
> > > > >
> > > > > Is this still the rule?
> > >
> > >
> > >
> > > _______________________________________________
> > > Chicken-hackers mailing list
> > > address@hidden
> > > http://lists.nongnu.org/mailman/listinfo/chicken-hackers