W3C home > Mailing lists > Public > xml-dist-app@w3.org > May 2001

[Repost] Re: Must understand mustUnderstand proposal

From: Jean-Jacques Moreau <moreau@crf.canon.fr>
Date: Wed, 16 May 2001 09:58:27 +0200
Message-ID: <3B023323.79A591D@crf.canon.fr>
To: Henrik Frystyk Nielsen <frystyk@microsoft.com>
CC: Williams Stuart <skw@hplb.hpl.hp.com>, XML Protocol Comments <xml-dist-app@w3.org>, Jean-Jacques Moreau <moreau@crf.canon.fr>
[Sorry about the previous mail; my mailer got confused.]

attached mail follows:

Jean-Jacques, I tend to agree with Henrik on this one. You are correct that
processed blocks are removed post-processing and so inter-block references
may break. However, it is unclear to me what you would expect an XMLP
processor to do in such situations? Checking references between blocks
would require the _XMLP_ processor to inspect the contents of blocks; would
you also have the XMLP processor check references from blocks to addresses
outside the message, and how would you deal with failure in such
situations? ...... I think this is a slippery slope that we should avoid.
I think this is an issue -- for application developers -- and we can help
them out by at least warning them of the problem: this warrants an extra
sentence in the spec, IMO.
BTW, is there a reason this discussion is taking place on private email? If
not, please move any reply over to dist-app.

David C. Fallside, IBM
Ext Ph: 530.477.7169
Int  Ph: 544.9665

                    Moreau"              To:     Henrik Frystyk Nielsen <henrikn@microsoft.com>                    
                    <moreau@crf.ca       cc:     David Fallside/Santa Teresa/IBM@IBMUS, Marc Hadley                
                    non.fr>               <marc.hadley@sun.com>                                                    
                                         Subject:     Re: [Fwd: Must understand mustUnderstand proposal]           
                    06:55 AM                                                                                       

Henrik Frystyk Nielsen wrote:

> Is this an issue that we have to deal with as part of the basic protocol
> or is it something that applications have to deal with? I would say the
> latter and I am not sure I can see what we can do about it without
> specific knowledge about the semantics of a header. In other words, I
> would say that this is out of scope.

Well, the current spec says that processed blocks should be removed from
messages. If the block that is removed was referenced by another block
within the same message, and that other block is processed at some later
node, the node will receive a broken message.

So, since we (the basic protocol) specifically ask the receiving end to
remove processed blocks, I would say that we have created an issue that we
have to deal with, or at least make people aware of.

Now, when we explore it, we may decide that we will leave it for phase 2.
(David, does W3C have any plans (yet) for phase 2?)

Received on Wednesday, 16 May 2001 03:58:50 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 22:01:13 UTC