The MovableType export format is still importable into WordPress with little trouble. I wrote a MovableType template to export all blog post titles and URLs so I could craft a .htaccess file to preserve old URLs, and that took a few minutes to work through reformatting from a plain dump to usable .htaccess structure with Redirect 301 directives. But not difficult, with the help of TextMate and regular expressions

The hardest part – by far – of the process was figuring out file locations, permissions, etc… on the target server. The migration was trivial once I worked that out.

Here’s the MT Template I used to generate the text that got processed into the .htaccess file for wordpress. There’s probably a better way, but this worked for me.

Looks like when I exported everything from Blosxom to the MovableType archive format for importing to WordPress, long posts got clipped. I’ll have to dig out the original source .txt files from the Blosxom incarnation of my weblog and fill in the blanks. What a PiTA.

It’s either going to have to wait for a while until things slow down a bit, or for a time when I need something banal and mundane to chew on while I try to figure something else out…

I’ve also set up an extremely simple page (actually, it’s a blog with no content of its own) that lists the last 20 posts on all blogs in the system. Works well, and I’ve got a crontask set to update it every 15 minutes so it should scale up nicely with a balance between up-to-the-minute posts and not hammering the system. It also provides an RSS feed for the whole system, which is pretty cool.

These are things that Drupal does as well, but in a rather more clunky manner, with more complexity on the screen. MT hides all complexity behind the Admin interface, so readers aren’t exposed to any of it (which is a Good Thing).

Personally, I prefer WordPress, and will continue to use it to power this weblog. But it has issues with scalability, making it less desirable for a large-scale multi-user multi-blog project on a campus scale.

It’s interesting that this non-project has gone from an idea tossed around between Paul and myself on the flight back from Northern Voice, to something attracting the attention of Information Resources, and of Information Technologies (who apparently are hesitant to come near this puppy for fear of being left to support it).

UPDATE: reworded some text to sound less confrontational.

Thanks to the heroic efforts of Brian’s team (especially Ken Pratt and Michelle Chua), they were able to resurrect the misbehaving database and migrate it safely to MySQL. Now, it should perform much better, and be easier to backup and repair if needed. I’m hoping Ken can be persuaded to document the process he concocted to help save future victims of the same problem.

Anyway, the weblogs at UBC are back on the air!

I’ve been trying this in fits and starts for a couple of weeks, and keep getting bitten by what appears to be corrupt data in the BerkeleyDB database. The import process (mt-db2sql.cgi) barfs on the Comments, Entry, and Theme tables. The import is rather useless without these three types of data…

Anyone out there have any magic juju to get at the data in BerkeleyDB? I just need to poke around in the database and look for corruption or invalid characters that might prevent importing into MySQL.

I see that BerkeleyDB is used by a whole lotta different software, but I haven’t been able to find a generic client application ala CocoaMySQL (which rocks for connecting to MySQL).

I wanted a way to have a single page with all of the latest posts sorted by date, from all weblogs.

RSS wouldn’t quite do it. It would sort first by weblog, then by date. Not what I needed.

Blagg + Blosxom sorta does what I wanted, but relies on creating cached text files from the various weblogs. Redundant, and icky.

Then, I tried the WholeSystem MovableType plugin. Works like a charm. I just created a new “meta” weblog, replaced all of the normal “blog post-ish” stuff from the template, and dropped in the WholeSystemEntries goodness.

Now, the index for the weblog goes out and grabs all posts from all weblogs on our system, and displays the newest 30, sorted by date, from whatever weblog they were from. Sweet.

Combined with the handy MT-Rebuild script, which can be set as a cron task, the Uber Weblog should be pretty self-sufficient. I may even add an Uber RSS feed – subscribe to it and you get the most recent LC weblog posts delivered to your aggregator in one easy feed… Might be a clean way to integrate this with the upcoming LC website update…

See for yourself…

If you run a MT weblog, you REALLY have to install this bad boy. Right now.

Here’s a screenshot of a small, small part of my MT log. This weasel tried several times over the course of a few hours, hammering repeatedly in bursts within seconds of each other (I can only assume it was some kind of software following Google links…)

Over the last 2 weeks, I’ve had to remove over 100 spam comments on this weblog. Over 99% of them were apparently from bot software that spiders the blog to bump up the Google ranking of the target site.

Anyway, long story short, I’ve installed this plugin (relatively painless), and it doesn’t require any modification to the MT code. Just drop 3 files into place, make sure you’ve got Storable.pm on your system, and configure away…

I’m crossing my fingers on this. I was |– THIS CLOSE –| to disabling all comments on this weblog because I’ve wasted too much time removing the unwelcomed spam (and much thanks to Yahoo! for doing SFA to stop their users from spreading this garbage.)

UPDATE: It’s been a week, and no comment spam! YAY!

It can take any RSS feed (static file, or URL), and parse it for display within a MovableType weblog. I’ve changed over the RSS subscriptions to CAREO on the right sidebar to use this plugin, and it works pretty well.

There are lots more great MT plugins at the MT-Plugins Directory.

UPDATE: I’ve also put the “other” way of subscribing to RSS feeds in a page here, since it’s valuable in contexts outside of MovableType (like, say, Blackboard, or WebCT, or whatever…)