So we worked out an arrangement to transfer the code and service from Tom to Automattic, and here we are. Here's what we've done so far over the past few days:
- We transferred the Rails application and most of the avatar serving to our WordPress.com infrastructure and servers.
- Avatar serving is now more than three times as fast, and works every time.
- We moved this blog from Mephisto to WordPress.
Wow. Talking about Rails bashing. Anyway, Om Malik is wrong about Gravatar -- it is not a "small project that gives Wordpress users the ability to add avatars to their profiles" -- in fact I use it in almost all my applications where a profile image needs to be placed, be that WordPress, Drupal or other custom built apps -- as long as a user can be identified by an email address. I guess I am just way too lazy to implement one myself, but at the same time, why reinvent the wheel when something as neat as Gravatar is available?
However Gravatar was no short of disasters. Their domain expired once, performance has been ranging from mediocre to very slooooow, and the coming of Gravatar 2 early this year did not seem to improve anything other than "hey, we are now running Rails behind Lighty 1.5beta!" That is why I coded the generic Gravatar Cache, as you really cannot count on it being available 24x7!
Now with Automattic powering the infrastructure -- the same company that gives us massively popular WordPress.com and Akismet services -- I can now sleep easy at night knowing Gravatar will continue to serve those 80x80 profile images even when my apps are dead. Oh wait, just let me check whether my sites are still alive...
What's more exciting is the plan for "Gravatar 3.0" -- although I suspect Automattic is going to absorb the Gravatar front-end into WordPress.com and keeping only the API, which is most useful for third party app developers anyway. That includes:
- CDN delivery
- Bigger image at 128x128px
- WordPress.com profile images accessible via Gravatar API