- From: Xiaoshu Wang <wangxiao@musc.edu>
- Date: Sat, 29 Sep 2007 11:11:33 +0100
- To: "Booth, David (HP Software - Boston)" <dbooth@hp.com>
- CC: Rhys Lewis <rhys@volantis.com>, Pat Hayes <phayes@ihmc.us>, Technical Architecture Group WG <www-tag@w3.org>
Booth, David (HP Software - Boston) wrote: > So I suppose if you do go down this route of defining a class of "HTTP > endpoints", which are the thingies that send 200 or 303 responses, then > you could define "information resource" as being the subclass of HTTP > endpoints that send 200 responses. > No. That would be a wrong definition. Assuming if I have a huge binary file at http://example.com/hugefile. On most request, I may redirect it to another document say http://example.com/hugefile.rdf, where I will do two things: (1) warn the requester that to retrieve the resource might take forever and (2) if they do insist on getting the resource, use a particular MIME type. So, http://example.com/hugefile is a "information resource" but can respond 303. Also, in reference to an earlier post of you that *why* it is important to distinguish "information resource" from non-IR. I think because IR are those things that we can directly manipulate with the SW technology and non-IR are those resource that we cannot. For instance, we can fetch an XML document, attach a XSLT stylesheet to transform it into another doc. But we cannot fetch a hard-copy XML document and shred them in pieces. Xiaoshu
Received on Saturday, 29 September 2007 10:12:10 UTC