I have been commenting for a while that I think that RSS item data needs to come with permalinks to itself, since most RSS is offered only temporarily through a feed with the most recent 10-25 posts.
With the Echo Project, I even suggested this kind of permalink be added to Echo-RSS itself. But, I also realized today this could be done consistently and conveniently without changing the RSS spec, if a slight change of practice could be made in the usage of the RSS autodiscovery links, which look like:
<link rel="alternate" type="application/rss+xml" title="RSS" href="http://icite.net/blog/?flavor=rss" />
The way these links are commonly used is to point to a RSS feed showing the most recent 10-25 items. Using a blog as an example, it makes sense that the "alternate" representation of a blog home page is a RSS feed showing the same 10-25 posts on the blog home page.
But, on a blog post permalink page, it doesn't make as much sense to indicate that the "alternate" view of that page is the 10-25 items on the blog home page. I am thinking that an "alternate" link on a webpage shouldn't be used to mean something other than an alternate view of that page.
So, my proposal is that, on permalink webpages, the RSS autodiscovery link be used consistently to indicate the "alternate" view of that page. For example, on the permalink page for this post, instead of using the same link I show above (which is for the RSS view of what is on my blog home page), I would use:
<link rel="alternate" type="application/rss+xml" title="RSS" href="http://icite.net/blog/200306/ rss_autodiscovery.html?flavor=rss&smm=y" />
This link is to the RSS version of this post, and it is itself a permalink and it contains the complete post in the RSS.
Because many or most blog tools could have permalinked RSS, this should be easy to implement at the blog template level. I know the blojsom engine I use to run this blog easily can do this, as I already have RSS permalinks that appear at the bottom of each post on its permalink webpage. So, blojsom can easily also produce the link tags I described.
For me personally, and with what I am doing with the iCite net presently, I see each RSS version of a post as a little database that can be queried and mined. I don't see RSS syndication and aggregation only happening in the scope of recent posts. Rather, I see it happening across all kinds of scopes including aggregation by topic, identity, profile, etc.
Having permalink access to structured data is, to me, practically just as good as having that data in a database ;-) In fact, because of the Internet, I think it is even better. That is one way to think about what the iCite net is looking to encompass.
So, what do you think about this RSS autodiscovery link usage change? If this seems like a good idea, write this up on your blog and spread the word. What should we do next to help link tags like this get commonly used?
Note: for more good information on RSS autodiscovery see Mark Pilgrim's Ultra-liberal RSS locator, which also links to additional posts with more background on the usage of the link tag.
Note: All comments and trackbacks are moderated. Spam is deleted. Other comments are approved as promptly as possible.
Note: Older posts no longer accept new comments or trackbacks.
Comment by: Joe Shelby · http://www.celticdistrict.com/joe.html
posted: Jun 27, 2003 4:34:14 PM
Question : Why is the mime-type "application/rss+xml" instead of "text/rss+xml"? I always thought that application usually meant binary data; XML is text, and should be typed as such so that a browser that doesn't have a handler for it at least has a fallback of displaying it (where the option to save is still there).