The man behind our main line of anti-spam defence is hanging up his hat.
Indeed, I am hereby officially announcing that I will no longer support, maintain or further develop Spam Karma (beside some very occasional, very limited poking, until the transition to a self-maintained project is completed).
Luckily he’s relicensed the code under the GPL, which means others may and can pick it up and run with it.
Why is he giving it away? Firstly because of the soul-sucking nature of engaging in an arms race with spammers:
Much as I love the challenge and excitement of coding an anti-spam filter and thinking up new tricks to defeat parasitic life-forms of the web, I just dont have the time anymore. And to be honest, if I did have the time, I probably would have other challenging, exciting new projects Id rather tackle. Im fickle like that.
I can certainly relate to that.
I can relate even more to the next reason:
I will really try to keep that one short, because I could probably write a novel of that. And it wouldnt be a very interesting read.
In a word: WordPress kinda sucks nowadays. Its retarded upgrade rate makes it nearly impossible to keep up, in turn making it a constant security threat on my servers. And each time I finally cave in and install one of those mandatory security upgrade, it also installs 600 Ko of other theme compatibility-breaking fluffy crap that I never asked for in the first place. Usually setting the ground for the next cycle of security-exploit-rushed-upgrade. To sum up, its become incredibly bloated and tedious to support. Replacing it on my own servers is very high on my list of things to do (which means somewhat in the first 1000 items).
Having no interest for WordPress anymore, I have thus very little interest for WordPress-related development.
Preach it brother.
The heart of the problem with WordPress is its release management strategy. Frankly, it sucks. They follow the Microsoft doctrine that bugfixes and feature additions belong in the same blob.
This means you have to upgrade when a new version comes out. And hope they didn’t introduce bugs that, say, drop the categories on hundreds of stories when upgrading. Because they don’t know if it will, because they don’t write tests to prove the software is sound, for the love of Mary.
A motto I have been kicking around is “Process is a Feature”, or perhaps “The Project is a Feature”. For an opensource project — indeed any software project — the way in which bugs are defined, features added, changes propagated etc is as much a feature as the tick-box list beloved of marketers.
If there’s an ironclad lesson that’s been learnt in almost every scenario, it’s that if you don’t have extensive testing and reviews, bug fixes deserve their own stream of activity, not to be rolled together with feature additions. One day the WordPress mob might click to this. But I’m not holding my breath.