The problem of the growth of syndication feeds

April 15, 2006

Sam Ruby has a problem: he has too many feeds. Specifically, he has a comment feed for every blog entry, and too many robots and abandoned aggregators are pulling them, over and over again, despite the feeds effectively being dead. (See his entry for his solution.)

Sam Ruby isn't alone in having this problem; anyone with a lot of flexible syndication feeds is going to encounter it sooner or later. (Even if you have only a few human readers, some search robots love syndication feeds. Even if you try to tell them to bugger off via nofollow et al.)

The more I've thought about this, the more I think that it's actually a social issue: we're so early in the syndication revolution that everyone is still figuring out how to organize things. Over time all of this will get solved; we'll work out what feeds are useful, users will form definite expectations of how sane feeds are organized, and there'll be 'good feed reader seal of approval' (social) standards for how polite feed readers should behave.

Some of this development will be technical, such as figuring out how to mark comments in feeds to clearly group conversation threads and link them to parent articles (there is a 'Feed Thread' draft RFC in progress, eg here). And anything that automatically mass-pulls syndication feeds is going to need to get much smarter about detecting overlapping feeds. (Failure to get smarter will result in mass lynchings.)

WanderingThoughts avoids some of these problems because I decided that the auto-discoverable feed for an entry should be the main feed rather than comments on the entry. While each entry does have a comments feed, people have to feed it to their aggregator more or less by hand, and the same holds for the various aggregated comments feeds. (In theory you can have several autodiscoverable feeds on a page, but I don't know if anyone handles that well yet.)

For annoying spiders that like to crawl my syndication feeds, I now use a blunt hammer; selected user agents get 403 responses on all requests for syndication feeds. I am optimistic that they will someday take the hint.

Written on 15 April 2006.
« Public interfaces and Solaris 9 patchadd exit codes
Weekly spam summary on April 15th, 2006 »

Page tools: View Source, Add Comment.
Login: Password:
Atom Syndication: Recent Comments.

Last modified: Sat Apr 15 20:04:47 2006
This dinky wiki is brought to you by the Insane Hackers Guild, Python sub-branch.