- From: Simon St.Laurent <simonstl@simonstl.com>
- Date: Mon, 02 Oct 2000 17:23:56 -0400
- To: James Snell <jmsnell@intesolv.com>, "'xml-dist-app@w3.org'" <xml-dist-app@w3.org>
At 01:54 PM 10/2/00 -0700, James Snell wrote: >Well... before we get too far along on this discussion, let's back up and >identify exactly WHY XML is currently NOT a good envelope for XML. I'm not actually concerned about why "XML is currently NOT a good envelope for XML." I'm not sure the 'envelope' approach actually makes sense any more, except for the sake of using legacy protocols - HTTP, SMTP, etc. While I understand the cost benefits of using those protocols, I have serious doubts about the wisdom of extending them in ways that will keep their limitations around for decades to come, and inflict those limitations on new work. There are some (Keith Moore, for instance) who see the work using HTTP as a substrate protocol as a threat to the integrity of HTTP. While I don't entirely agree, I do see HTTP as imposing potentially serious limitations on the structure and potential of future protocols. For something small like XML-RPC, I don't think this matters very much, but it feels like a serious risk for larger and more complex protocols. HTTP 1.1 uses keep-alive to create a kind of envelope for efficiency reasons, but I suspect we need something a little more flexible over the long term if we want to support large-scale flexible information transfers. Simon St.Laurent XML Elements of Style / XML: A Primer, 2nd Ed. XHTML: Migrating Toward XML http://www.simonstl.com - XML essays and books
Received on Monday, 2 October 2000 17:20:44 UTC