Wednesday, August 06, 2003


On combing RSS with RDF: This has been going back and forth quite a bit in the last week. Here's something to think about: While the idea of richer properties embedded in an RSS flow makes sense in a general way, I am not sure of the job listings is a good application.

It seems to me (newbie that I am) that the essence of RSS is the idea of a time ordered flow of read-only information, where the recipient is most interested in looking at the newest ones first, and where the information is timely, with a short shelf life, and generally somewhat informal.

I don't think a feed of job listings fits that at all. A job is posted and remains posted until it is removed or filled, or the definition changes. You want to easily look at the current list of openings. You want to easily see which ones are gone. None of this sounds like an RSS feed. Similarly, an RSS feed of recipes or telephone numbers doesn't fit in my mind.

If you agree with that model, then brainstorm with me on what kinds of information would fit into that model beyond articles. Here is some off the top of my head ideas: traffic reports, weather reports, snapshots, last minute tickets (e.g. theatre).


4:09:14 PM  >  trackback []   comment []       


More on RSS Auto Discovery. Here are some useful links:

  1. BlogStreet has an XML/RPC API which includes a call which will return the RSS URL given a blog's URL.
  2. There's been some very recent activity towards standardizing an HTTP header field pointing to the corresponding RSS URL, thereby eliminating the need to 'auto discover' by simply putting the answer in the header.

10:56:50 AM  >  trackback []   comment []       


David Weinberger is still a genius.

Sorry I am just testing to see if my TrackBack macro is working for me, and it doesn't appear to be. I will trace my debugging of it here:

  1. Initial attempt: added the macro per instructions to my template. It visually looks correct but always indicates "0" trackbacks.
  2. I note the "URLs to Ping" box. It's not clear to me what that is exactly. I would have assumed that any blog URL in a post would be pinged automatically. I am now guessing this is not the case.
  3. I edit the post and manually ad the URL to David's post that I am testing against in that box. This still doesn't seem to work.
  4. I wonder what URL is required: is it the permalink or the root of the blog or what? I assume it's the permalink, which of course is hard to find in David Weinbergers' post which I am testing against. I wander around and find that the date of the post is actually a link to it's perma link. Who knew? Still it doesn't work: I see a zero both in my perma link count and in his. I would expect to see a zero in mine but a one in his.
  5. Further research and I locate a post from when this feature was in 'beta' where it is stated that the pinging only happens when I initially add the post, and not when I edit and post changes (like I am doing now.) Not knowing if that matters, I create a new post, fill out the URLs to ping box. Still doesn't work.
  6. I re-read the spec and found that in fact it is not necessary to place the URL of the target post in the URLs to Ping. So that was a dead-end assumption. But it still doesn't work.

8:11:52 AM  >  trackback []   comment []