A little bit of remodeling
by Josh Staiger
I'm redoing a few things in my site, so don't be alarmed if you suddenly see some funky color schemes.
Firstly, I have renamed my weblog from "Josh's Weblog~" to "Josh Staiger's Weblog~". This is all part of my coop attempt to become the number one ranked page on Google for a search of "Staiger". Damn you Staiger-Fahrrad! You will fall yet! (And yes, I realize that my linking to them is not helping the cause, but it's good for morale) The real reason for this change is that I think it makes my site title a bit more distinguished in RSS readers, Technorati, Feedster, etc...
Secondly, I have upgraded to Movable Type 3.11. I was mainly hoping to take advantage of the new dynamic templating system, but it now appears that I won't be able to do that until I start using MySQL as my database on the backend. That will have to wait for another day for now, but I'm excited about the possibility of dynamic templates as it will make tweaking the appearance of my page much less of a pain in the ass.
Speaking of the appearance of my page, I have fixed the individual entry page so it no longer displays comments in a completely retarded manner. However, this fix is causing other CSS weirdness. Hence the all-white background for the moment until I have time to look into the problem.
On the topic of comments, even though it is entirely unsurprising, I am quite unhappy with the way that the number of comments I receive has taken a serious dive since I started using requiring a valid TypeKey profile. The good news is that I just noticed that MT-Blacklist is finally available for Movable Type 3.1 (another reason why I wanted to switch to 3.11). The bad news is that it is still in beta and on attempting to try it, it spewed forth all kinds of nasty error messages. I spent a few hours trying to hack it to get it to work, but to no avail. This, too, will have to wait for now.
I'm still curious to try a bayesian filtering solution for comments and there is such a plugin available for MT, but as the author of the plugin clearly states there are many problems with this approach at present, namely that unless you have a very high-traffic weblog, it is very difficult to get a sufficient volume of comments (both good and bad) to adequately train the system. There are some obvious solutions to this, I think, such as having a central spam corpus repository (or tapping into such extensive collections that may already exist for email). More things to think about...