Re: "Art Of Unix Programming" on Perl

[Edward Peschko]

> Language design is a very tough nut to crack, and we decided (as
> a group) that we don't want a language designed by committee, we
> want a languaged designed by Larry.  

Right, but does it hurt to give general guide-posts on how the language is to 
operate? If everybody knows that it is going to be microkernel based and 
have pluggable parsers, why NOT announce it to the world? Put it in stone.

After all, larry himself said that he was trying to break it up into sections 
and announce those separately...

The best we can do (frustrating
> as it may be) is to let him think deeply.

True, but then that's what PR people are for - to get PR generated. How do you
expect *anyone* to get that job done if he/she doesn't have access to the people
that would make positive PR possible?

I'm all for letting 'larry think deeply' but then again, I'm all for having him
check in with the real world every once in a while. That's a large part of what
a PR job is *for*.

> > That's a problem in itself, because if people don't see progress they lose
> > interest and go away.
> 
> Perhaps.  If they go away to hack Perl5, is that a problem?  If they
> go away to hack Python or Ruby, were they really _that_ interested in
> Perl6 in the first place?  And is it really a problem if the lack
> of interest magically disappears when Larry returns?

Well, that's debatable, isn't it. Perl has a certain level of 
rubber-band-eability, as has its community. Its tested that limit before, 
it could test it again.

Personally I'd like to see new-blood in the development community. I don't want
them to go to other projects. A lot of productive work could be done if some 
decisions are made; we don't have to be delivered perl6 on a plate 'in toto'.
That's not what design is about.

> Rushing the process because of intermittent PR problems isn't going
> to make Perl6 any better at achieving it's goal - solving tomorrow's
> problems better than Perl5.

Again, I don't think that this would be rushing things at all. Give a couple of 
RFC's to chew on, announce it to the world, develop interest, give the 'perl
internals' people stuff to work on, generate discussion, move from high-level
design to medium-level design. 

Ed