- From: Elliotte Harold <elharo@metalab.unc.edu>
- Date: Mon, 26 Feb 2007 10:32:44 -0500
- To: Jan Algermissen <algermissen1971@mac.com>
- CC: noah_mendelsohn@us.ibm.com, www-tag@w3.org
Jan Algermissen wrote: > Not sure if this misses the point of you proposal, but it might be a good idea > to also mention that self-descriptiveness of messages is what protects the > sender of a message from an effect of change the receiver might go through. > > Messages that are not self-descriptive depend in part on implementation of > the receiver. In an environment where there is no cenral control that keeps > the receiver from changing the client would never be able to independently > prove what it actually said when it sent a message. This seems grounded in the idea that the sender's meaning and the recipient's are the same, and that it's a problem if they're not. The sender sends what it sends. The receiver interprets that in the way that is useful to the receiver, irrespective of what the sender meant. Adding more or less markup doesn't materially change that. The sender can prove what it sent, if it cares to. However absent a contract with the receiver, it has no expectation of or claim on particular processing by the recipient. Similarly, absent a contract or agreement, the recipient cannot insist on any particular meaning or message format by the sender. The title is "self-describing documents" but what people are writing and talking about are more "self-agreeing documents" and I find that a very dangerous concept for several reasons. The sender's mere description does not imply or mandate any agreement or processing on the part of the recipient. -- Elliotte Rusty Harold elharo@metalab.unc.edu Java I/O 2nd Edition Just Published! http://www.cafeaulait.org/books/javaio2/ http://www.amazon.com/exec/obidos/ISBN=0596527500/ref=nosim/cafeaulaitA/
Received on Monday, 26 February 2007 15:32:54 UTC