- From: Arthur van Hoff <avh@marimba.com>
- Date: Thu, 11 Sep 1997 11:03:41 -0700
- To: Push Workshop <www-push@w3.org>, DRP Mailing List <drp@marimba.com>
- Message-ID: <3418327D.57DBAA52@marimba.com>
Hi, Thank you all for a great workshop. As promised I would start a discussion on this mailing list to decide which working groups should be assigned to work on the DRP protocol. Before we can decide this we need to determine the overlap/synergy with other ongoing work in the W3C and IETF. There are three major areas of functionality in the DRP proposal that are fairly distinct, and that may be discussed seperately: 1) Index Format - There seemed to be a feeling that there was some overlap with the Web-DAV definition of collections. Can someone from the Web-DAV group please provide a pointer to the relevant documents which define Web-DAV collections so that we can determine how closely related they really are? 2) Content Identifiers - There is clearly a need for globally unique and verifyable identifiers for content. Work has been been done in this area, but we need to have a discussion to define content identifiers more precisely. Here are some relevant sources of information: - URN: http://www.ietf.org/html.charters/urn-charter.html - LIFN: http://www.netlib.org/utk/projects/rcds/rcds-tr/node6.html 3) HTTP Protocol Extentions - The DRP proposal in its current form uses some new HTTP headers. The Content-ID header is similar but slightly different from ETag. The Differential-ID header is new and has interesting new sematics which enable the download of only file differences. Maybe Jim Gettys can suggest ways on how to proceed. Although more than one working group may be involved in further work on DRP, it seems that there should be a seperate working group formed which is tasked to define interoperable transport solutions for push applications using HTTP and Multi-cast. The first task of this group would be to finish a proposal for a DRP based protocol using HTTP, followed by further work focussed around the use of Multi-cast as a transport. Thoughts? Input? Ideas? Have fun, Arthur van Hoff
Received on Thursday, 11 September 1997 14:04:34 UTC