livejournal status update
brad at danga.com
Thu Jun 23 18:58:25 PDT 2005
Over the past two days David Recordon and I have been busting out the
LiveJournal UI and LJ::OpenID changes necessary for leaving blog comments
using OpenID, from a previously unauthenticated state.
This is an improvement on our status of a month ago where we could comment
using OpenID only after we'd logged in, which is by far the rare use case.
Usually [non-LJ] people will stumble upon LiveJournal from elsewhere, and
immediately want to comment, not go login first.
So what we do now is, on submit, store a "pending comment" (which gets a
number), send the user for their OpenID auth loop, and when they come
back, finish posting their comment. One of our regular maintenance tasks
keeps the pending comments table clean, giving users a more than
reasonable amount of time to finish the auth loop.
All very straight-forward, but I'm excited about it because this afternoon
the DeadJournal guy (Frank) updated his beta-site with the new code, and
David was able to leave comments back and forth from LJ's beta site and
DeadJournal's beta site. This has been on our wishlist for 3-4 years now.
Probably as of tomorrow we'll update the production LiveJournal.com server
code and DSA support will be gone. LifeWiki.net will be updated to use
the new consumer library.
In other news:
MovableType server support is mostly done, and today I saw one of the Six
Apart developers add OpenID consumer support to MovableType.
TypeKey/TypePad would've gotten OpenID server support in the next couple
days, but it looks like with the protocol changes it'll be pushed out a
couple weeks to sync with one of their dev cycles. TypePad consumer
support development hasn't started, but given that the MovableType
consumer support was done in an afternoon, I'd guess the next
version of TypePad will support it.
We have a Google Summer of Code intern wanting to do OpenID support for
Python, Ruby, MediaWiki, and WordPress. (hopefully we get him! :-))
Um, that's about it.
I'll send out mail again when LiveJournal.com is running the new server
code. We'll also be pushing the LiveJournal consumer code, but we have a
global switch to keep it disabled for the first couple hours/days.
 http://changelog.livejournal.com/2005/06/22/ ,
More information about the yadis