It's an unrewarding and long slog.

Probably because most of the Perl6 design is done in secret. First the language, and now the compiler. No one contributes because, well, no one really knows what is going on besides the inner cabal. I guess they feel that they can do a better job without the rest of us interfering. I emailed Patrick right after he was announced as compiler pumpking offering to help, and I was told that they want to have a prototype done before they even take it to the list. It doesn't feel very "community-rewrite"-ish or even "open-source"-ish to me.

Now, contrast that with parrot...

I think you're inadvertently mischaracterizing Patrick's comments.

Now, contrast that with parrot...

Parrot's a project several years further along, used in production, with tens of contributors, and a lead designer who prefers to design the skeleton of a system before announcing it and showing off code. Aside from the first three, they're not too different (and the first three are a function of time).

There's really no point in distributing a couple of empty header and source files right now. How do you solicit patches from volunteers if there's nothing to patch? Also, the new lists aren't ready yet, so there's a good reason not to announce things on the wrong list.

That's not to say that there's no discussion; see p6i from just today with comments on regex semantics from Larry, Dan, and Patrick.

Yeah, but the point with parrot is that the lead designer has kept the design *open* since the beginning. Sure, there isn't any code yet for the real compiler or real grammar engine. However, the design is still happening behind doors. There have been something like a half-dozen prototype grammar engines over the last four-or-so years done by many different people, and that's a lot of input that's being shut out.

The design of Perl 6 was started with RFC's. I cannot imagine how this can be considered secret.

Ok, and after that process was over, how has Perl6 been designed?

First, a few points

• the community isn't rewriting the compiler, they are rewriting the language
• the design of the language has never been secret
• it's hard to open-source something that has no source yet

Do you begrudge Patrick and Luke some private time to get their act together? You'd rather that happen in public? Do you fear that they will head off in the wrong direction or that they will miss something vitally important before publicly releasing code?

Or are you afraid they'll waste time covering ground that's already been covered (as I glean from another post in this thread)? I see this as Patrick's time to ramp-up. Even if he covers already-covered ground that's okay because he's the one charged with doing the work and he needs to learn the things he needs to learn to get the job done.

duff

Definitely point B. As I stated in the node that you're referring to, there's been a bunch of Perl6-related grammar engines over the last few years. For instance, take a look at this list, and then notice that the date was over 2 years ago. Since then, we've had: Perl6::Rules, japhy's Perl 6 Regex Parser attempt, Steve Fink's latest attempt, and now Patrick and Luke's attempt.

Don't get me wrong, from what know from talking with Patrick a few weeks ago I definitely think he's on the right track. I have no doubt that he'll do a great job, but its just... *frustrating.* Almost every one of the dozen or so grammar engine attempts so far have been a solo or semi-solo act, and every single one is only partially completed. Many of them were just prototypes or toys, so that's ok, but this is going to be the real thing. I'm just a little worried that the same thing will happen. (that we'll have Yet Another Unfinished Engine)