- From: Greg Stein <gstein@lyra.org>
- Date: Mon, 24 May 1999 16:56:35 -0700
- To: jamsden@us.ibm.com
- CC: w3c-dist-auth@w3.org
jamsden@us.ibm.com wrote: > ... > suggest we add a BATCH command to HTTP/WebDAV that has an entity request body > consisting of an XML document where each element is a WebDAV method, its > attributes are the method headers, and its content is the entity request body if > any. The mutlistatus response could be extended to include and entity request > body element for each returned status if any. This would be generally useful, > would reduce communication round trips, and would eliminated the need to specify > anything special for MKCOL (which hasn't been specified yet). It would also > provide a simple transaction mechanism if the BATCH method was atomic. Vendors > could advertise canned BATCH commands that provide higher-level functions in > much the same way we are experimenting with user-defined property sets. Any > thoughts? I believe this won't buy you anything over HTTP/1.1 connection pipelining. You'll still transfer just as much information, and you'll still use a single TCP connection. About the only thing that this could do would provide for the potential for atomicity. IMO, that isn't enough of a win to compare to this additional complexity. Cheers, -g -- Greg Stein, http://www.lyra.org/
Received on Monday, 24 May 1999 20:04:46 UTC