Alas, I'll be doing very little Perl for my new job; lots of C/C++, and my new overlords seem to have fallen pretty hard for Guido's koolaid, so I'll need to write a Python indenter . While I'll inject Perl where I can, I fear it'll be an uphill battle, and avoiding those is why I succumbed to captivity in the first place.
OTOH, I'll be doing stuff I've been doing for the better part of 20+ years, stuff I can do without getting out of bed, much less falling out. And I have a few notions I hope to inject into the product suite; the database field is starting to get crowded again, and anything we can do to differentiate is going to be essential to success.
And they're letting me do it from Reno, which was a big condition for me. The California thing lost its luster for me years ago; I especially revile the SillyConJob Valley nonsense. I live 10 miles from the CA line...and thats plenty close enough for me.
I also made them pick up the tab for my OSCON attendance, so I'll actually get to stay at a decent hotel this year!
What really made me steer back into the database maelstrom was the crew manning this startup. Some of them are almost iconic in the DBMS world, and they've been very succesful in the past; some others I've personally known for a decade or more. Hopefully, the cumulative experience will be enough to climb over the other startups massing around this particular niche. I commented to the VP of Engineering that, if the DBMS s/w didn't pan out, he could just pimp us all out as consultants at $300/hr and probably make more money!
In the meantime, I'm desperately trying to get Thread::Sociable into at least compilable shape and uploaded to a Google Code repository I setup some months ago. Unfortunately, life keeps getting in the way, so progress has been very slow. Its a damn shame I can't get paid to finish that up. I really think it would make a nice splash for Perl: faster shared access, real STM for real threads, and even shared nothing for those that prefer it, plus (for those that hate on ithreads) all that for process shared memory (without the Storable or DBM contortions of existing solutions). Hopefully, some others will attempt to grok my XS/C hacking and take up the cause. While I won't be tossing commit bits around like beads at Mardi Gras, I'll certainly open things up for those that express both an interest *and* the skill to comprehend it (the STMness gets pretty gruesome in places).
Oh well. Its my last weekend of independence, and another beautiful Sierra spring day, so I'd better get out and enjoy it before I jump on a plane on Monday.