Ok that's fine but can we include a recommendation to that effect in the spec, i.e. just after we say that the sequence is not determined, we should say that if you want a defined sequence, don't use a compound request. Stephen. Rich Salz wrote: > > Stephen, > > At the call today we decided to explicitly leave requests within a > compound request as unordered. That is, a server is free to respond to > the requests in any order. If sequencing is important to a client, it > should make multiple synchronous requests, for example. > > We couldn't understand a good use case that justified the additional > implementation burden on servers. > > /r$ >Received on Tuesday, 9 November 2004 17:46:33 GMT
This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 20 September 2007 14:31:02 GMT