Summary of current state of httpRedirections-57

My thoughts on ISSUE-57 (httpRedirections-57):

When this issue was opened the following were on the table:

1. Bug in HTTP spec where it says 303 response MUST NOT be cached,
   leading to inefficiency in semantic web nose-following.
   -- Everyone acknowledges that this was just a mistake in RFC
   2616. It has already been fixed in HTTPbis.

2. Need for a non-3xx response so that the original URI stays in
   the status bar for bookmarking.
   -- Could be 203, some new 20x, or a 4xx
   -- No discussion on this since 2007 as far as I can tell

3. Rhys was working on a finding that was a followon to the
   httpRange-14 decision (basically deployment advice), but his work
   stopped with the advent of the SWEO IG note ("Cool URIs for the
   Semantic Web").
   -- This work needs to continue at some point, as technical
   decisions are supposed to eventually find their way into
   architectural recommendations.

Some new business started to be tracked under this issue:

4. Link: and "uniform access to metadata" were tracked under this
   issue, but subsequently moved to ISSUE-62.  (As the TAG made a
   decision that Link: was OK for some purposes, this is on
   track to incorporation into an arch. rec.)

5. As a response to some work on redirections by David Booth, Tim
   wants to do something with Tabulator.  This has to do with
   inferring that the 200 after the 303 is a description of the
   original resource and so on. (I'm not sure what this has to do
   with the TAG.)

Current business that is not being officially tracked but is ongoing:

6. During this spring's TAG priorities discussion, Jonathan and Henry
   said they wanted to see the resource / information resource debate
   driven to a conclusion.

My personal recommendations:

1. Bug - solved by HTTPbis, we're done.

2. 203 - refer the concerned parties to the HTTPbis process,
   which has been amazingly supportive of our GET/303
   advice.

3. Finding - leave ISSUE-57 open to track work on the necessary
   finding.  I am considering volunteering to take on the task
   (maybe because my name is a misspelling of Rhys?).

4. Metadata - we are on course with ISSUE-62

5. RDF - on course with AWWSW.  But we should think about what
    problem(s) AWWSW is solving and what issue(s) to track
    it under.

6. IR - this would be forced by the process of creating a
   finding (due to intimate connection between httpRange-14
   and IR)

-Jonathan

Received on Monday, 15 June 2009 15:44:55 UTC