News Aggregators Most Groovy

I was really intrigued by Steve Gillmor's piece here in EWEEK about Groove and especially liked what Ray Ozzie said about news aggregators vis-a-vis their new release.

Q: We've often talked about the Groove opportunity to integrate RSS aggregation and routing features. Where does that stand with Version 3?

We learned a lot with the RSS aggregator work that we're working on in-house. It's very easy for us to bring sets of feeds together from multiple interested people to look at. The only part that I don't think we've nailed yet is what happens when the useful information can only be seen when you're not looking at the summary—when you click the links, do you then have to go out to the site? People haven't packaged those sites in a way that they can be taken offline.

When I look at your site feed, the URLs in the summary point at HTML pages; they don't point at MHTs (Multipurpose Internet Mail Extension HTML [MHTML] format), self-contained blobs that I can pull down. I would have to do a deep crawl of your site in order to pull down all the content into that shared space. Unfortunately, the Web standards people have not done a good job yet at, [is that when] given a URL, how to write a method to pull down the MHTs.

Q: The model I'm looking for here is the ability to capture an RSS feed that contains the full text or an XHTML rendering of it. And it would also solve some other significant issues related to RSS aggregation and authoring—namely, multiple machines, where Groove file-sharing enters the picture.

What we've done in Groove—not in the packaged product, but we've definitely done a few versions of it [that] our customers have used—is Groove aggregation. That same Groove aggregation technology works across your machines. If you read something on one machine, it gets marked read on the others. And it can aggregate multiple feeds, and so on. There are so many aggregators out there. I don't know whose is going to ultimately gain the most market share. But we have done a bunch of experimentation on that, and I look forward to, at some point, either we or one of our partners releasing a good group aggregator. That's a pretty obvious, useful application.