Loading ...
Sorry, an error occurred while loading the content.

114Re: [RSS2-Support] another namespace question

Expand Messages
  • Jake Savin
    Oct 1, 2002
    • 0 Attachment
      On Monday, September 30, 2002, at 10:06 PM, Phil Ringnalda wrote:

      > Some parts of Dublin Core (date, creator, contributor, and subject) are
      > quite widely used. But using dc:description rather than description
      > is, um,
      > unusual. With my feeble understanding of RDF, it seems a bit backward,
      > since
      > the RSS 1.0 schema [1] defines description as a subPropertyOf
      > dc:description, so I think the way the theory goes is that an RDF
      > parser
      > that doesn't know from RSS will find and parse the RDF schema for RSS,
      > note
      > that description is a subPropertyOf dc:description, and then when
      > anyone
      > asks it for a dc:description of Infoworld, it'll say "here, here's an
      > RSS
      > description, that's just a particular type of dc:description." I didn't
      > think it worked in reverse (if I say "gimme a fruit", you say "here's
      > an
      > orange", but if I say "gimme an orange", you don't say "here's a bowl
      > of
      > cherries, grapes, and pineapple"), so I would expect that a
      > schema-aware RDF
      > parser would say that feed doesn't have an RSS description.

      That's what I would think on the face of it too, but the fact that I
      saw only <dc:description>'s in the InfoWorld feed, without also having
      <rss:description>'s (or <description>'s) threw me for a loop.

      In terms of current practice, it seems as though at least some
      non-namespace-aware aggregators will happily present a <dc:description>
      to the user, even though it may not be aware of the Dublin Core at all.
      Is this the case? Is this also the case for namespace-aware parsers?

      Or -- should we consider this to be a bug in InfoWorld's feed?

      > But, as Morbus said, practically, you just add everything you've ever
      > seen
      > to a list of possibilities, in descending order of preference. Given
      > that
      > Radio users are more used to seeing full items with HTML, I'd say
      > content:encoded [2], description, dc:description. For a start.

      That sounds like a slippery slope to me, unless I'm misunderstanding
      (which is of course quite possible). ;-)

      It seems to me that either I understand an element or I don't. Saying
      that I understand an element not intended for me (in this case
      <dc:description>), but only if the one I was looking for
      (<rss:description>) is not present seems wrong.

      I don't want the addition of namespace awareness to Radio's aggregator
      break our users. I also want Radio to do the right thing when it
      encounters an element it doesn't know about.

      Looks like my choices are:

      1) I should leave it the way it is, and live with the possibility that
      some feeds will display content to Radio users that was not intended
      for them.

      2) I have to add support for the Dublin Core to Radio's aggregator.
      (Are there others? Which ones?)

      3) I should report a bug to InfoWorld and live with the breakage after
      adding namespace support to Radio's aggregator.

      Am I interpreting the issues correctly, or did I miss something?

      -Jake
    • Show all 12 messages in this topic