W3C home > Mailing lists > Public > xml-dist-app@w3.org > November 2003

Re: Proposed resolution for issue 440

From: John J. Barton <John_Barton@hpl.hp.com>
Date: Fri, 07 Nov 2003 09:39:20 -0800
Message-Id: <5.1.1.6.2.20031107092417.01fc1888@hplex1.hpl.hp.com>
To: "Herve Ruellan" <herve.ruellan@crf.canon.fr>
Cc: Martin Gudgin <mgudgin@microsoft.com>, "Xml-Dist-App@W3. Org" <xml-dist-app@w3.org>

Thanks Herve; this sounds ok but I want to explore it a bit more to be 
convinced.
I appreciate your patience. Please see below.

At 10:05 AM 11/6/2003 +0100, Herve Ruellan wrote:

>John,
>
>See my comment below.
>
>John J. Barton wrote:
>
><snip/>
>>To put this in use-case terms, consider a multipage UI carried in a message,
>>each page consisting of the same large template image combined with a bit of
>>unique text.  Why force the template to be sent multiple times?
>>Or am I misunderstanding: is there another layer of indirection here?
>
>Yes, to prevent the transmission of the template multiple times, you can 
>add another layer of indirection in the XML message. In this way, each UI 
>page would refer to an XML element which would contains the large template 
>image.
>
>In its current state, MTOM only does *some* optimization for message 
>transmission, mainly by transmitting base64 encoded data directly in binary.
>It could be argued that MTOM could do more optimization such as checking 
>that it sends the same base64 encoded data only once.
>However, I think it is better to do this optimization at a higher level: 
>in your use case, the creator of the multipage UI knows that the template 
>image is the same for each page, therefore, the creator should take some 
>steps to insure that this template image is include only once  in the XML 
>document. This has several advantages:
>- The creator of the XML document will probably have efficient way of 
>knowing that two template images are identicals, while the MTOM level will 
>probably have to compare those images bit by bit.
>- The creator of the XML document may not know in advance that this 
>document will be transmitted by a "special MTOM" able to detect and 
>optimize duplicate binary data.

In reality the application that creates the document and the application 
that packages it for transmission will be different.  For example we will 
have something like Dreamweaver to create the content and something like 
infozip+ftp to send it.  The created XML document will have references to 
external binary files, probably as relative links.  A file may be linked 
multiple times.  I understand the process under SwA: I pick up the SOAP 
root, find all the relative links,  assign them to ContentLocation values 
in my package, create the MIME, zip, send.  Under MTOM I am not sure how to 
take those first steps since the original information model isn't a tree. I 
guess I have to create an XML representation of my file system directory 
with entries being "includes", re-write the relative links in the SOAP root 
to point into this XML data, then pass it to MTOM?

Thanks,
John.



>Best regards,
>
>Hervé.
>

______________________________________________________
John J. Barton          email:  John_Barton@hpl.hp.com
http://www.hpl.hp.com/personal/John_Barton/index.htm
MS 1U-17  Hewlett-Packard Labs
1501 Page Mill Road              phone: (650)-236-2888
Palo Alto CA  94304-1126         FAX:   (650)-857-5100
Received on Friday, 7 November 2003 12:39:32 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:59:15 GMT