Breadth can be good. And unless your market is different from Silicon Valley, being as much as a year out of date on a particular stable technology isn't that big of a negative. (Being out of date on an unstable technology is kind of iffy, and things in Microsoft land look to be getting unstable again as people sort out .NET.)

Something to ask yourself is whether you want to compete in the market based on your tool skills or based on your problem-solving skills. Problem-solving skills have a longer half-life than tools skills, and tend to get you placed higher in projects. When tapping architects for a project, I look to thinking ability and exposure before looking at narrow technical skills.

If you do decide to market yourself based on tool skills, there's some benefit to having a rare skill, as long as there's at least some market for it. Plenty of people know C, fewer know Perl. Fewer still know how to speak Perl, SQL, and HTML. That's a skill set that I'd bet on, were I going strictly technical. Databases are hardly going away, and HTML will be with us for a while. Perl is a great way to tie the two together.