Peter Seibel wrote: > I think in the *very* long run the idea of a "Community Lisp" > is an interesting one (and a great name).
Thanks :-). > But I also think it > needs to be approached very carefully. For better or worse, > the current standard is the locus of a lot of strong > feelings--on the one hand it is one of the great strengths of > Common Lisp and it was also produced at great cost, both > financial and emotional. So any proposal to muck with it in > any way is--fairly or not--going to stir up a lot of > strong emotions. But that's just it - I'm NOT proposing to muck with it. ANSI Common Lisp will remain ANSI Common Lisp, and will probably be set in stone as THE ANSI lisp standard since I doubt the J13 committee will be gaining enough paying members to take any major action. If a NEW standard (or standards even) emerges, it's a strictly voluntary thing and is unlikely to ever get the ANSI stamp. It will be VERY similar to ANSI Common Lisp, since it will build off of it, and virtually all of what goes into the new system will likely be things added in addition to the current work. I'm not proposing to rip apart anything. The only changes to the Common Lisp foundation would be things that are very widely agreed to be mistakes (I think the ANSI tests Paul Dietz has been putting together have shown a few things that definitely need clarification or fixing, and are agreed upon virtually universally.) The main focus would be things that really NEED to be standardized (e.g. a universal FFI) but aren't right now. > While eventually it might be desirable to get the dpANS > licensed in a way that would allow the creation of an open > source derived work, that seems to me to also be putting the > cart before the horse. Not really, for two reasons. a) Just because the status of the spec is clarified doesn't mean we HAVE to work on building something new from it. Once it is clarfied what we can do with it we have the option, and there is no disadvantage to having that option as soon as possible. b) Finding the necessary people to talk to is not going to be a simple matter, and could take time. By the time there is actually a need for clarfication due to external reasons, clarification may be all but impossible. Ten years is bad enough - imagine what fifty years would be like. I suppose I shouldn't have mentioned the Community Lisp idea, since it implies that's the sole reason for getting the status of the draft clarified. It's not, by any means - just being able to distribute dpANS3 with a lisp distribution would be a benefit. There are many good reasons to have matters clear. > It seems to me better (and more in > lines with the CL Gardener philosophy) to work on building > grass roots support for whatever things might go into a new > language specification and *then* worry about how to get > those things codified. I'm not worried about how to get them codified, I'm worrying about not being able to use the draft spec when the times comes TO codify them. As you say, that's for the future. But there are reasons for clarifying the spec status now, even if it will be a long time before we need it. > The whole point of having a standard, de jure or de facto, > is to encourage conversations like this: [snip] > But the important thing to remember is that there's > nothing about having even an ANSI standard that forces the > conversation to go that way. The implementor can also come > back with, "Yeah, technically you're right but I think that's > an unimportant part of the standard." Sure. But that's an argument about making a new standard, not clearing up the draft spec status, which is all I'm really talking about at this time. The Community Lisp idea was just an illustration of the potential benefits - there are others if that one doesn't appeal. > Thus the only thing that *really* causes the conversation > to go the way we want is that enough users care about feature > FOO that enough implementors decide they better get with the > program and implement it rather than blowing it off. That's usually application driven - if users are actually USING things do do interesting stuff, it's a definite motivation. Camm has been great about this on the GCL side of things. > The best way, I think, to build a strong base of user > support for some new feature is to go ahead and build it-- > either as a portable library (c.f. Portable Common Loops) or > as a proof of concept in some particular implementation. The > former seems more likely to succeed than the latter if only > because the latter tends to trigger the not-invented-here > antibodies in other implementations unless the new feature > is so incredibly and obviously good that everybody just has > to have it. And even then, folks never seem to be able to > resist mucking with things. Another possibility, and one that > can proceed in parallel, is for folks to start now > contributing to the various open source Common Lisp > implementations so that when the time is ripe for the > Community Lisp revolution, each open source Common Lisp > implementation has at least one member in good standing who > supports the idea. The reason the ANSI standardization worked > at all was because the people involved were the implementors- > they were in a position to agree to implement the standard > they wrote. Sure. But that's for the future. If we clarify the status of the draft ANSI spec NOW we will be in an infinitely better position to take action in the future. > So that all said, I'd be much more interested in building a > list of things that might differentiate Community Lisp from > Common Lisp and then looking at how to get each of those made > available across Lisp implementations *without* having to > write a whole new language standard. That's the whole point of making the draft spec status explicit - so we won't have to write a whole new language standard. I'm not saying we need to do ANYTHING with Community Lisp now, or within any definite time frame. For now, I just want to get the draft spec status clarified. That's it. I was trying to illustrate motives for taking the effort doing so, but those are nebulous right now - the one concrete proposal I'm making is that we get organized about clarifying the status of the draft spec. Cheers, CY __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com _______________________________________________ Gardeners mailing list [email protected] http://www.lispniks.com/mailman/listinfo/gardeners
