Unsurprisingly, things haven't stopped being busy. No real development happening in the open source arena.
I've decided to drop techbot completely. I'm not even going to kid myself that I'll ever complete that project, and it's not a world-changer project anyway. Instead, I'm going to devote my spare hacking time to writing a small but flexible business management system, based on various 21st century technologies, and release it open-source with the hope of scratching a personal itch, and helping small companies that can't afford big SAP-like packages get started with a free, open source management system that they don't have to pay extortionate prices to tailor. I think it's a big hole in the GNU/GNOME portfolio that I'd be interested in filling. I've been designing it in my head for years, and even developed portions of it, so it's about time to put it all together and come up with something a bit more concrete. Hopefully, once I've got priority one (work) out of the way, this'll be my priority in Thailand. I'm sure the local Thai companies would find it useful, and might help me with the internationalisation. Watch this space :)
Recently, I spent a couple of days replacing my co-hosted server with something a bit pokier, and redesigned the mail backend to use Messagewall/Postfix/Cyrus IMAP, authenticating against a MySQL database, with a simple, but easily extendable web PHP4/XML/XSLT frontend (http://www.golder.org/). With any luck, that should now run itself quite happily for some time to come without needing too much intervention by me.
The other day, I spent the day rigging up my home network with IPv6, and learning as much as I could about it. I got as far as being able to setup a tunnel, and a routing advertisement daemon, so machines behind my IPv4 gateway were able to ping the 6bone happily. It's all interesting stuff, but I've still got a way to go before I really understand it properly. Next time I play with it, I'll be looking into how to set up automatic site-local addresses.
Basically, work projects have been taking up most of my time. I'm a bit of a workaholic. I love my company and have done for years before I came to work there. I've always felt strongly about the companies I've worked for, and it always upsets me if I feel things are going in the wrong direction. Nildram are definitely going in the right direction, it's just going to take a while to get there :)
My work projects are great though. They're all based on open source products, and I love working on them. In fact, our latest project involves a lovely technology combo - PHP4/XML/XSLT. Trouble is, I've found I've been getting distracted from my software development responsibilities recently. We've got a bigger problem to deal with - our mailservers.
Whilst I'm not officially supposed to get involved in Operations, I always do. I'm on their mailling list, to pick up on any software related issues that arise, so I'm normally in tune with their common problems.
At the moment, the problem is that our main mailserver, hosting several thousand customer POP3 accounts, was built up over many years, using patched versions of software (I helped fix up some of the patch, but it's still a bad solution), and working off address/routing databases that have suffered from bitrot over the years. Over the last year, matters have deteriorated, with the server regularly going down due to various unfortunate hardware and load problems (exacerbated by it's poor configuration and susceptibility to spam attacks). The problem is really starting to upset customers, who are upsetting our support guys, who are looking to Operations for a solution. As nobody else had stepped up with a solution, I articulated a whole load of problems which I consider to be contributing factors, and posted an action plan to address each problem in turn. The end result wil be that all the addresses we've 'lost track' of will be tracked, and we'll end up with a more efficient, more spam-resistant mail system. Customer's are happy, support are happy, everyone's happy, right?
Wrong. My boss is understandably a bit apprehensive and concerned about the possibility of problems during this upheaval. Whilst I can't say there will be no problems, the potential for it solving a number of our longer-standing support and customer satisfaction problems is immense. I've spent a large amount of my time in this job analysing and fixing mail-related problems caused by this unusual configuration. Now that there is a decent plan of action to clear up all our known configuration issues, plus the possibility of migrating to a newer faster servre, running a more efficient (unpatched) MTA, now are exciting times.
Unfortunately, in all this I've learned that sometimes people will just carry on trying to work around a problem day in, day out, rather than just tackle the problem properly and move on to better things. It reminds me of a comedy sketch about a DIY enthusiast whose house fell down.
Still, just 8 weeks left, then the hammock should smooth away all my work-related worries, and my lack of involvement in day-to-day network operations should mean I get to concentrate on coding.