I spent the morning doing the "search" track at SXSW. First was Don Turnbull's panel on Revolutionary Search Technologies, followed by Marissa Meyer talking about Google.
I asked a question of the first panel, more related to information architecture than search. "Since RSS has become a fad, people have been rushing to move all their structured data into it—even if it's not an article for syndication. Is that a good thing?"
Scott Johnson fielded the question. He said a year ago he would have said stuffing everything into RSS was a bad thing. He's changed his mind. People are putting all sorts of things into RSS and it seems to be working. At the very least, it gets the information partitioned properly.
After the session, I introduced myself to Scott and explained that I was thinking about this problem in the context of event scheduling. I'd like to see organizations, say, around Austin be able to publish their event calendars in an XML format. Then people could aggregate a calendar of events from whichever organizations they might choose. RSS provides a powerful syndication mechanism for the project, but not the appropriate structure for schedules. Scott pointed out there is an RSS extension out there for scheduling. I was excited to discover that. This project suddenly seems more feasible.
I tagged along with the search people for lunch. Over ribs and chopped beef sandwiches, I mentioned one of my worst, current RSS annoyances. I stuffed the Austin Bloggers blogroll into an RSS feed, but it won't validate. The <author> tags for the blog entries sometimes are names or 'nyms and not full email addresses, which violates the RSS spec. I guess I'm not the first person to complain about this. Sam Ruby rolled his eyes and said, "Use <dc:creator>. It's a perfectly good tag." I'm not sure why I overlooked the Dublin Core tags, but that's the perfect solution to my problem.