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

Re: ETags as first class objects

Expand Messages
  • griffinc18
    ... they have a ... would they ... a HEAD ... I ve never seen it either, hence my question. It seems if you have ETags as URIs that could be dereferenced,
    Message 1 of 3 , Aug 23 1:35 PM
    • 0 Attachment
      --- In rest-discuss@yahoogroups.com, "Sebastien Lambla" <seb@...> wrote:
      >
      > I've never seen ETags used with Uri values.
      >
      > What would be the semantics of inspecting the ETag? Just because
      they have a
      > Uri doesn't mean they can be dereferenced, and if they are, what
      would they
      > return? A description of a specific version through http headers in
      a HEAD
      > request or a content type describing lock and update semantics?

      I've never seen it either, hence my question. It seems if you have
      ETags as URIs that could be dereferenced, then that means you're
      keeping around old versions of your resources to return. Or I suppose
      you could 303 old ETags URIs to the latest one, but that requires some
      sort of ETag history.


      > There's also mention of the possibility of the server doing a merge
      on the
      > fly from a put. Isn't this in conflict of the work being put into
      > resurfacing PATCH?

      I suppose, but remember, this document was written in 1999. Has there
      been much progress on PATCH, BTW? I can't seem to find anything recent.

      Thanks,
      Griffin Caprio
      312.371.3869
      http://www.griffincaprio.com
      http://blog.1530technologies.com
      http://www.oreillynet.com/xml/blog/


      > -----Original Message-----
      > From: rest-discuss@yahoogroups.com
      [mailto:rest-discuss@yahoogroups.com] On
      > Behalf Of griffinc18
      > Sent: 23 August 2008 05:50
      > To: rest-discuss@yahoogroups.com
      > Subject: [rest-discuss] ETags as first class objects
      >
      > Hi all,
      >
      > I was reading http://www.w3.org/1999/04/Editing/, which discusses
      > using ETags to combat the problem of the lost updates. It makes
      > mention that the authors believe that a bug in the design of ETags is
      > that they are not first class objects. In this case, first class
      > objects means making them URIs.
      >
      > I was wondering if anyone else has heard of something like this or has
      > seen it implemented anywhere? I've never encountered ETags when they
      > were anything other than values in HTTP headers.
      >
      > Thanks,
      > Griffin Caprio
      > http://www.griffincaprio.com
      > http://www.oreillynet.com/xml/blog/
      >
      >
      >
      >
      >
      >
      >
      >
      > ------------------------------------
      >
      > Yahoo! Groups Links
      >
    Your message has been successfully submitted and would be delivered to recipients shortly.