So Google has blog search. Summary: It’s fast, it’s reasonably complete, it’s stripped-down in the typical Google style, the result ranking needs work, the time window is way too deep. They’re also providing feeds, which is good, but the feeds are horribly buggy [Quick response; One big bug’s already fixed!].
Why Feeds? · I actually don’t find blog search that interesting. There are a few times where I wonder “What are people saying today about lemur propagation”, but usually not. Since by “blog search” we really mean “feed search”, and feeds are by definition a stream of data flowing by at high speed and volume, what’s way more interesting is to have a few canned queries that you hold up against the stream and have the results trickle out as a feed when they happen.
Well, Google, supports this, which is good. They support RSS and Atom, which is good. But the “Atom” is version 0.3, which is completely dead, has no official standing, isn’t being maintained, is being thrown on the floor by the feed validator, and has no real advantages. Please, Google, can we get with the program?
The other problem is that the feeds are broken.
[Update: No, they’re not any more; the
entry/id values were
silly but now they’re fine.]
It’s real easy to forgive early-beta-release bugs, particularly when they get
fixed like this; who’d want to distribute software any more?
But there’s still no excuse for shipping Atom 0.3.