« iPod for News | Main | On Sitemaps »

Monday, February 02, 2009


Feed You can follow this conversation by subscribing to the comment feed for this post.

Sean Gillies

Erik, do you think a study of RSS or Atom formatted sitemaps would be an interesting follow-on? Atom was designed, in part, for such applications.


And then there is Google's proposal for geographic data sitemaps. That could also be interesting to explore.



@sean: thanks for the comment! the current ways in which sitemaps.org supports feeds is very minimalistic. essentially, they just read any feed and treat it as a source of information for URIs on the site. not all that exciting, it's basically just feed-based crawling.

we are mostly interested in how a sitemap could actually become a "map". currently, "sitemap" really is misleading because it's not a map, it's just a list of links. we are interested to really support a map as a navigational structure of a site (what people probably expect when they simply hear the term "sitemap"), and sitemaps.org do not support this.

what's worse, the current version of sitemaps.org is rather loosely defined and extensibility is not clearly defined at all, which makes it hard to design extensions. a new version of sitemaps.org should be out soon, but i don't know how it will work. but it should be more well-defined in terms of how to extend the sitemaps format.

wrt google's geo sitemaps: that's really interesting! i did not know that one. and they're extending sitemaps to do that. but like i said, it's currently hard to find out about the upcoming extensibility clarifications, and even though the current format is "open", google so far has adopted the usual kreml-style communications strategy they have around anything that's not officially released: do not admit anything!

Sean Gillies

You're saying the "sitemap" is really more gazetteer than map? I'd agree with that.


@sean: exactly. i would argue that anything that claims to be a "map" of something should put these things into some structure. the current sitemaps format has absolutely no structure for the URIs; it has a very small amount of optional metadata, and apart from that, the URIs all just go into the big bucket of URIs for a site.

The comments to this entry are closed.