Via hublog: "NCBI's PubMed is to offer RSS feeds for searches. Only took them 2½ years."
Update: Pubmed RSS is now live. After doing a Pubmed search, click on the drop down menu for "send to" and RSS is now one of the options. The feature allows you to specify a name for the feed and number of items to include, see here for an example of the result. Pubmed is using RSS version 2.0 and dumping the content of the HTML citation display directly into the description tag. For most users this won't matter as the display in most aggregators, like bloglines for example (see the bioinformatics folder), will be the same as the regular Pubmed site. However using RSS 1.0 with available modules would have enabled more metadata to be added to the feeds, see for example the semantically rich feeds produced by Connotea. Which is great if you're a Semantic Web fan, which we all are right ?
Interestingly it looks like RSS export is going to be part of the Eutils web services (see the feed url). Now if RSS 1.0 export was available from all the NCBI databases then we would have instant RDF data.
I'll leave commenting on issues such as the correct identifiers to use in the Pubmed RSS (lsids ?) and whether or not the RSS 2.0 feeds are valid, use best practices (entity encoding the HTML in the description elements) etc. for later as I'm busy with ISMB related activities.
Comments
RDF
I've recently added RSS 1.0 feeds to HubMed - here's an example feed.
You can also export RDF from abstract display pages, for use with tools like Piggy-Bank (which picks up the links automatically) - here's an example of one of those items.
Piggy-Bank
Excellent. At the moment I can't think of a killer demonstration of why RDF is a good thing for biology, however RSS 1.0 feeds from pubmed and abstract RDF export are a good start.
I tried Piggy-Bank with hublog RDF export as soon as I saw your post about it on hublog. I eventually had to nuke the extension as it was starting to cause Firefox to slow down to a crawl.
As an idle thought I'm still not sure I'm keen to have metadata itself as the focus of an application. Do I use Piggy-bank to bookmark, collect citation data, sequence data ? Or special purpose applications that also produce metadata ?
I guess I should also put my money where my mouth is and start adding tagging, RSS 1.0 and RDF dumps of posts on nodalpoint...
Piggy-Bank
Yes, Piggy-Bank could be useful if it didn't have that bug (which might actually be a Firefox bug) that destroys your browsing within 10 minutes. The developers know about it, anyway.
I'd say use Piggy-Bank to bookmark for now. Hopefully, you'll be storing the data as well as the metadata in most cases.
I don't know...
It's been almost 20 days since that post, haven't seen any changes yet. They didn't give a specific time frame, just said 'real soon now'.