Thus far the posts on this subject (at least, those I have taken the
time to scan/read) concentrate on the issue from the publisher's
perspective. Is there a description anywhere of how this change will
impact on clients for this data?
For example, in order to be able to access Linked Data resources in the
context of XSLT processing, I have written a CGI "forwarder" which takes
a target URL and a desired content type as parameters. With the help of
this forwarder I can use the standard document() function to retrieve
RDF/XML resources relating to any target URL, so long as it is served up
in a manner which follows the existing 303 strategy. There are two
points about this client strategy which might be impacted by the current
proposals:
1. In order to work with the standard document() function, the forwarder
sets the HandleRedirects property of the TidHTTP object (this is
Delphi-land) to true. This means that it has no knowledge of whether
the initial response was a 303 or a 200. Will the client be expected to
know whether a redirect has happened, and adjust their strategy accordingly?
2. The 303 strategy allows the client to specify the preferred response
type in the Accept header, and in my implementation this is used to
request RDF/XML, plain XML or even Topic Maps. Will the revised
strategy support this same degree of control over the format of the
response?
Thanks,
Richard
--
*Richard Light*