Over the weekend I knocked up an OPML browser. Still not quite sure what it's for or if it's useful.

Things I noted mostly to do with validation against Dave's proposals.

- OPML not served with an XML content type so that the browser doesn't know how to display it.

- Wholesale abuse of the Type attribute. There's lots of OPML out there with a type of LINK pointing at RSS.

- I could do with a list of all possible values of TYPE. Is there only LINK and RSS?

- Confusion over the use of URL, XMLURL and HTMLURL. Most of the OPML I've looked at uses only URL even for feeds.

- The lack of a HEAD.DESCRIPTION element.

- All the same issues as in RSS around the inclusion of HTML in the TITLE and DESCRIPTION elements.

- Confusion over the use of the TITLE vs TEXT elements. They seem to be interchangeable.

- Using the file extension to identify the target URL type as another OPML file seems old fashioned. Particularly where the target URL is so rarely a static file but much more likely to be a RESTful directory or CGI string. I think it would be more useful to be explicit and use a TYPE attribute of OPML to tell the reader what's at the other end.

Overall the OPML standard badly needs some more concise guidelines, validation and for people to go round and complain when well known OPML feeds don't follow them.

[ << A Last.FM for books. ] [ Let's do Web 2.0 >> ]
[ 17-Oct-05 3:29pm ] [ ]