John Tropea has some wishes on the opml-front:
- A search tool for searching a specific opm-file: I like the idea, but there is a risk when such a tool also searches in the opml-files that it finds within the opml (inclusion). Searching full-text across a massive directory sounds great but it is like asking a search-engine to search on a specific page and also search on all the pages it finds links of on the first page, and so on… What if opml-file #1 links to opml-file #2 and #2 links to #1. How do we get out of the loop?
- Subscribe to an OPML or get services to subscribe to your OPML: I agree, if you are developing a service that can handle opml, make it handle it dynamically and not just import the feeds!
- OPML archive or latest posts for your blog: In my opinion oplm is not such a great format for storing posts. From a developers perspective storing html in an attribute instead of between tags where you can escape the html with [CDATA[]], is a bad idea. Al the angled brackets have to be replaced by < and > and some more. And why would we do this anyway when we have a fantastic format that was designed for storing posts: RSS. Grazr and bitty will handle this very well!