Looks like you have a space for it(?). If I were going to try to do a cert scheme, and I've considered it before, I'd probably break it into pieces, theoretical parts, tests, reqs, areas of expertise. BrowserUK for example can hack the pants off of me, as it were, in huge tracts of land problem domains but if you're hiring for CGI/Ajax I would be a much better choice because he doesn't do, or care to, that stuff. So any realistic cert plan would have to be, like some good Perl, modular.
Then I'd come back here with one standalone piece at at a time for feedback. The critiques might be brutal but they're also usually solid. Whatever survives, or is at least not rendered entirely impotent, could go into the scheme. I don't like design by committee, you usually get the lowest common denominator or an ongoing process that takes... um, when did Jon Orwant throw that coffee cup again? Critique by the meritocracy would save many cycles of putting it together; again, if it can be. It would be a lot of work with very little love in return.
The main reason I'm pro-certification despite agreeing with the criticisms is the publicity, visibility, and discussion it would generate. The sense of Perl being a legitimate, first-class programming language shouldn't be an epiphany that is reached by the odd, above average tech manager. An entirely worthless certificate might benefit the larger Perl world just by existing, even if it wasted some time for the average hacker or put extra pointless paper on the HR desk. Plenty of awful Perl hackers are getting hired every day without certificates.
(update: fixed spelling of Jon.)
