- From: Stefan Eissing <stefan.eissing@greenbytes.de>
- Date: Mon, 3 Feb 2003 11:49:59 +0100
- To: Jim Luther <luther.j@apple.com>
- Cc: w3c-dist-auth@w3c.org
Am Freitag, 31.01.03, um 18:28 Uhr (Europe/Berlin) schrieb Jim Luther: > > I thought it was pretty clear in rfc2518 (and rfc2518 bis) that WebDAV > servers and clients are HTTP/1.1 servers and clients because the > introduction section of the rfc starts off, "This document describes > an extension to the HTTP/1.1 protocol that allows clients to perform > remote web content authoring operations." I agree that WebDAV server and clients should implement HTTP/1.1. > However, during interoperability testing I've found several servers > which return HTTP/1.0 in the Status-Line of their response message and > behave as HTTP/1.0 servers (i.e., they do not handle HTTP/1.1 > persistent connections correctly, there are no Content-Length headers > in a response with a message-body, etc.). For the most part, the Mac > OS X WebDAV file system client works with these servers but it works > much slower because of transaction aborts and retries. So, servers implementing 1.1 are rewarded with better performance/user perception. Good. > > Should WebDAV clients refuse to use HTTP/1.0 servers and should WebDAV > servers refuse to work with HTTP/1.0 clients? I think refusing to interop with a HTTP/1.0 server is not the intention of HTTP/1.1. Contrary, HTTP/1.1 was carefully designed to allow interoperability. Even if all servers support HTTP/1.1, there are a lot of poxies wich only support HTTP/1.0. SAP's implementation will therefore keep support for HTTP/1.0. > > If yes, should rfc2518 bis make HTTP/1.1 (or later) a MUST? > I think it does already, maybe it should say so in a stronger way. But I doubt that it will have the consequences for your production code that you'd like to have. //Stefan
Received on Monday, 3 February 2003 05:50:52 UTC