It’s always sad to see a good project go a direction that you’re not going to follow it in. Case in point: Power Phlogger. It’s a neat stats application that I was always partial to because it gave information like user resolution and color depth that you don’t usually get from stats programs. It is called through javascript or a backup “image” so you only get “people” in your stats, not bots. It makes it easy to view your stats in terms of actual people visiting your site, what pages they went to, how long they spent on each, information that I find a lot more useful than “X number of people came to your site in May, here are the browsers they used.” Most web logs (not weblogs) have a lot of redundant information that can be easily abstracted in a relational database. (Okay, weblogs too.)

However it’s been half a year since the application has been updated, and much longer since there have been any significant upgrades. They basically stopped working on it to focus on PowerPhlogger3, which is going to be built from an entirely new codebase. That should have been my first warning. Part of the reason WordPress has been so successful when other PHP blogging applications that started about the same time haven’t is that it built on the b2 codebase rather than rewriting everything from scratch. The old code had a lot of problems, but it’s something we’re improving incrementally with each version. (The old code also did a lot of things right.) What if the Firefox developers had decided they needed an entirely new rendering engine and we had to wait 3-4 years for the first release of Firefox? The release date for PPhlogger has fallen back again and again, and no code is currently available to the public. It went from requiring PHP 4.2 to not working on anything but PHP 5, which hasn’t even been released yet and is a long way from being available on most hosts. Along the way they created yet another PHP5 framework. Whenever version 3 comes out it will run on a dozen different databases (11 more than I need). Everything is object-oriented now.

I’m sure all of this is very exciting from some sort of computer science standpoint of code purity, but on the other end there is an impatient user. The situation is made worse by the fact that, as I have found on wordpress.org, PPhlogger 2 does not scale well, to the point of slowing down everything else on the server. I ended up just removing it. I’m going to have to turn it off on this site soon. To some extent logs become useless when your traffic grows; you just can’t watch stats like you used to. That’s why services like Technorati are popular amoung high-traffic bloggers—they extracts meaningful data (who’s linking to me?) out of the noise of web stats. I’m looking for another program that will do this.

PowerPhlogger’s first release candidate will come out “no earlier than July 2004,” a date that has been moved several times. I hope I am wrong in thinking the project has jumped the over-architected shark and they release an amazing product that is fast, useful, and stable.