So, I finally get round to starting an Iterative Software journal.
We've just waved Leon off on his trip to Washington for a wedding. We don't think he's going to tell anyone about Graphing Perl while he's over there...
Meanwhile, I'm trying to work out how to turn stealth marketing into real marketing. The sort of thing that has people thinking "Now that looks like the kind of service I could use".
This is hard. No, this is very hard. It's especially hard when you don't have piles of VC funding behind you to help pay for advertising (not that advertising necessarily helps, but at least you feel like you're doing something). Writing this journal entry probably counts as displacement activity to avoid having to think about marketing.
At least we've got the website in a vaguely finished state, though it's crying out for a redesign, and we've got a couple of training courses available. Must sort out the course dates.
In my programming life, I'm busy trying to get my head around lenzo's infobot code. I discovered that the non POE version of the code falls over in a big heap if you're not very particular about which database format you tie to (and I've got the horrible feeling that it could fall over in a slightly smaller heap in some edge cases when it's not using the more fragile database libraries.)
It seems that DB_File and Tie::RDBM, both of which are wonderful, don't like it if you fork while something is tied via them. What happens is, the child process does its thing, dies, and DESTROY does some tidying up. But that tidying up confuses the state of the file handle back in the parent process. And badness ensues. Which isn't ideal.
All the more reason to use POE, I hear you all say.
Well, yes. But geckobot currently doesn't talk to IRC. And it hurts my head (big time) trying to suss out how it works. I feel a massive hacking session coming on.
