The TOCRoSS project was a JISC project carried out last year in conjunction with Talis, Emerald and the University of Derby. The main goals of the project were to look at how RSS feeds could be extended to include sufficient metadata for populating library OPAC systems, thereby allowing users to search and then access more content from within the library environment.

The project write-up goes into more details and provides links to some sample RSS feeds and a (yet to be populated) Sourceforge project which will be home to some of the code developed during the prototype.

While the experiment was no doubt worthwhile, I think the project is something of a missed opportunity, as it failed to take into account any of the existing work thats been done on enriching publication feeds with detailed metadata.

For example Nature, Ingenta, and other publishers have been producing RSS feeds containing rich metadata for several years now. Collectively we've converged on RSS 1.0 supplemented with Dublin Core and PRISM metadata to carry precisely the information added to the TOCRoSS feeds, only using RSS 2.0 and ONIX. You can compare and contrast examples.

The existing formats have already been successfully used to drive other services, e.g. CiteULike, so there's demonstrable implementation experience that suggests that they meet a number of use cases.

While there's certainly plenty of room for debate about which format(s) and vocabularies are better -- perhaps Atom should replace all of them? -- its a shame that the project didn't focus on drawing out the existing lessons learnt and areas for further work.

Also, had the project adopted existing formats then the code that will ultimately be published would also have been much more useful: it would immediately work with thousands of existing feeds, as opposed to requiring publishers to support another variant format.