Quick Thoughts on WordPress 3.0 Beta

A few days ago I installed the current beta release for WordPress 3.0 (WP3), which is scheduled for general release in the next few weeks. I wanted to see what the “gotchas” might be as well as get a head start on getting my themes up to snuff. Here are some of my impressions with getting things up and running.

Since a huge attraction for me with WP3 is the merging of its codebase with that of the multi-blog version, WordPress MU, I needed to tweak my development system to allow for wildcard domains. With that done, I created a database and did the usual WordPress install.

In order to get the multiuser aspect running, you have to tweak some code in wp-config.php. Though it’s not onerous, my hope is that this is a bit more automatic with the final release.

I then discovered to get the new menu system working, you needed to add a function into the theme’s function.php file. This isn’t a problem when using the default “Twenty Ten” theme, but will be something to pay attention to for pre-WP3 themes. In particular, URLs will have to be changed to reflect that uploads (images, music, etc.) are no longer in the “/files” folder but are instead in the “/files” folder. Again, not hard to deal with.

There will be some tweaking to get everything “just so”. Some plugins can be activated for all sites, some need to be site-activated. It’s things like that which will take up some time initially, but isn’t much more of a hassle than getting a WPx site up for the first time.

There are some bugs about (it’s a beta, after all), but so far I’ve only discovered one item that will cause site admins to gnash teeth, pull hair, and other ways to vent their vexation: migrating users from pre-WP3 blogs. I’m still playing with it, but after many hours I’ve yet to have a site transfer successfully. Even when I can avoid a PRIMARY KEY error in MySQL, the site(s) simply won’t run correctly, if at all once user data is transfered.

As many sites don’t want to lose their long-gathered user bases, this is a very significant stumbling block. There needs to be a relatively easy procedure to not only migrate user rolls from various blogs into a multi-site WP3 install, but also to have them correctly assigned to the correct blogs. The migration needs to preserve emails, web URLs, and passwords. Not a lot of administrators will stick with WP3 if they can’t maintain their existing user roles. Obviously, this is what I’m most working on at the moment.

Except for the user-migration thing, it looks like WP3 will be a winner. I especially think the point-release will be enjoyed by most.