Yup, I changed it to make your eyes bleed signify my newfound self-acceptance (and renamed the blog accordingly). Those colors were supposed to be washed-out shades of red and purplish-blue, but I was told on #linuxchix that the header looks pink. Sounds weird to me — I wasn’t specifically trying to make it look “girly”. If it’s pink, then it’s a shade I like — I’m not a fan of the shades presented in the Wikipedia infobox.

Igor Gomes drew my attention to the fact that Launchpad OpenIDs didn’t work in my blog. The OpenID implementation used here was self-written and only supported OpenID 1.0, so it’s no wonder — but still, when I looked at the replacement that I used, the PHP OpenID Library together with this component, it made me weep. OpenID 1.0 was elegant in its simplicity, and so was the authenticator code; is OpenID 2.0 bloated?

Still, Launchpad userpage URLs work in the OpenID box now. EAUT is supported too, so you can use an email address if you have one linked to OpenID.

There’s now a search box, and the “archive” link on the sidebar works properly now.

The new blog software is gradually coming to life. I’ve got tags and RSS working, plus moved some articles to Textile and added redirects from their old URLs. I didn’t map WordPress categories to new tags because I didn’t see the point – it will be better for me to later tag every post manually, since I’ll add some new tags anyway.

Per-tag RSS redirects from the old address for the purposes of Planet Ubuntu, but global RSS doesn’t. Therefore, if you’re using RSS on this site, you should update your client to use the new URL as linked from the sidebar.

Two pages are also available at new addresses (with redirects from the old ones):

• The Expanded Otomeverse
• Time Travel in Stargate SG-1

The “Mai-Otome GEMs” article, originally saved from Wikipedia, is now available at this address on Wikia’s Mai-HiME Wiki.

To convert HTML (as used by WordPress) to Textile (as used by the new software), I use this Python script, originally downloaded from here and tweaked to read HTML from stdin and write Textile to stdout.

To determine what legacy URLs need to be mapped, I use the public inbound links page. What, you didn’t expect it to be public? Well, both the QDB and Sideways Spa contain one each as well, and admins can see your IPs. Be warned.

What needs to be done:

• Archive
• Tag cloud
• More articles

You know, for a native English speaker, it’s probably hard to take a company named “Softage” seriously, even if it is Russian.

Amazingly, a certain Scottish company does…

Regarding the blog: a lot of stuff still isn’t working. This includes the “archive” and “rss” links, plus proper pagination (the main page is currently set to display the last 20 posts). Comments work, but old comments still need to be imported from WordPress — and since many of them are anonymous and new comments are OpenID-only, I’ll need another hack field…

Edit: And pagination works now. I’ll try to deploy this pagination to the QDB this evening, to replace the awfully hackish one it currently has.

The new blog has been deployed.

And it’s a disaster so far. WordPress keeps posts and pages in the same table, which has resulted in many pages turning into posts.

But at least, Dorothy, we’re not on WordPress anymore. Maybe one day I’ll find or write a converted from WordPress’s fubar HTML to Textile, so that I no longer need to have a separate hack flag to display the old posts…

Edit: Copied the wpautop function from the WordPress source code to process old posts. Indeed, the only way to process WordPress posts is to do it like WordPress does… as opposed to an open format, such as Textile.