Google has released “Sitemaps” which is yet-another-format for doing “site summary” or site-metadata publication. Unlike commonly used RSS formats or Atom, Sitemaps are only intended to communicate information about URLs that have been updated – not the content of those URLs. However, it seems quite clear to me that even though the goals of Sitemaps are a distinct subset of those we have for Atom, there isn’t really much reason why they needed to go off and define a completely different and incompatible format.

 

Perhaps if this group had dealt with the issue of “pinging” this divergence wouldn’t have occurred. (Yes, I know it wasn’t in-charter…) The Sitemap format can be thought of as very much like a “ping file” since it basically contains little more than info on the update time for urls.

 

I’ve blogged about SiteMaps at:

http://bobwyman.pubsub.com/main/2005/06/google_sitemaps.html

 

There are two formats defined by Sitemaps:

1.       A Sitemap format which lists URLs and provides their last-update time, priority, and update-frequency.

2.       A Sitemap index that points to other sitemaps. (One sitemap is required for each unique path-tree on a site.)

 

Something similar to Sitemaps would actually be quite useful in helping us work out the issues related to sites that have multiple feeds (i.e. three flavors of RSS and one or two flavors of Atom…). One might provide a Sitemap Index that listed each feed. This would make identifying duplicate feeds much easier than it is today. A Sitemap might also be a really interesting place to put information on a sites signing and encryption keys…

 

Perhaps we can convince Google to work with the Atom Working Group on starting an effort to define more compatible formats and also to build a pinging standard that would be useful to both the blogosphere and the traditional web. We’ve proven in the blogosphere that providing site-metadata and summary information is very useful and we’ve proven the value of pinging. Now, it will be wonderful if we can help move those methods to the rest of the web.

 

            bob wyman

 

Reply via email to