[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Y Store now C++

Dorai Sitaram wrote:

>>Well, that's why Common Lisp was like that. We did not want to tell
>>all our users that they had to run a tool over all the programs that would
>>somehow try to find all those name conflicts and fix them: they would
>>have screamed bloody murder.
>I take it you mean that this approach was actually
>considered and rejected?  Hm, I didn't think it was
>considered, not because the people concerned wouldn't
>have thought of it, but because they must have actively
>preferred Lisp2 over 1.
The way I remember it (Guy can perhaps correct me if I'm wrong?), there
was seriously great pressure from the Lisp1 people, who were utterly
grossed out by Lisp2 which they viewed as a holdover from the dinosaur
period and a disgusting wart on the language. Again, the way I remember it
is that Moon and I were pretty much in sympathy with this but felt that the
compatibility problem was just insurmountable. My memory is that this was
one of the single most contentious issues standing in the way of getting a
concensus between the parties that were trying to get together and agree
on a Common Lisp. So, of course, we tried every avenue of approach we
could think of to figure out what compromises we could make, whether we
could somehow have it both ways, and so on. If the old common-lisp
mailing list is archived somewhere, I bet you would find a huge number of
messages about this.

>  Cf Pascal Costanza's
>recent question on why one should switch to Lisp1?  How
>would you answer someone like him so that he would be
Lisp2 means that all kinds of language primitives have to exist in two 
or be parameterizable as to whether they are talking about the value cell
or function cell. It makes the language bigger, and that's bad in and of 

>I don't see this source transformation so unworkable at
>all... if there was will behind it. 
Ah, but the question is, whose will? It's one thing for Moon and me to be
sypathetic to the langauge design principles of the Lisp1 people putting
pressure on us. It's quite another thing to go to the users, who are people
working hard trying to get a job done, and say, hey guys, we're about
to pull the rug out from under you, you have to run this tool over all your
source code, you have to worry about co-ordinating that change with your
other changes, and so on. If you've ever managed large software projects
you will appreciate the tremendous difficulty of coping with an incompatible
langauge change, tools or no tools. We're talking very large software bases