Re: Poll (Was: Question for Implementors (Was: Schema Validation Transform))

Speaking just as a WG member, Option 1 leads to the longest delays in
achieving higher standards levels.  The key phrase is "wait for

Option 2 seems a bit flakey. There is a risk the higher approval
bodies will decide this is really just a variation of 1 and impose
delay. A better subvariation of option 2 would be to move them to an
informational Appendex but see 3 / 4 below.

Option 3 (labeled 2) and 4 seem best from the point of view of
advancing the XMLDSIG spec. And if we think we know what we want to
say about these transforms, which I believe we do, then putting them
in the additional URIs/algorithms informational document seems like it
adds value.

So I think option 4 is the best way to go.


From:  Joseph Reagle <>
Organization:  W3C
To:  "Gregor Karlinger" <>,
            "merlin" <>, <>,
            <>, <>,
            <>, <>, <>
Date:  Wed, 19 Sep 2001 17:38:42 -0400
Cc:  "XMLSigWG" <>, "Eastlake" <>
References:  <>
In-Reply-To:  <>
Message-Id:  <>

>On Tuesday 18 September 2001 05:33, Gregor Karlinger wrote:
>> we (IAIK) have not yet implemented XML and schema validation transform.
>My rough tally then is 5 implementors have responded that they do not have 
>immediate plans to implement XML or Schema validation as a Signature 
>transform. Baltimore has some support for both [1,2]. This is fine, no one 
>is advocating these features as requirements. Folks will get to them in due 
>time. However, our problem is that folks *will* get to them, and they'll 
>wonder how to do it properly. This question has already identified a few 
>ambiguities in our spec that we've been able to fix.
>The immediate question facing us then is what to do with these parts of the 
>spec in the mean time? Please send your response (particularly from 
>implementors) by the end this week. Should we:
>1. Retain the sections [3] as is and wait for interop.
>2. Retain the sections  [3]in a modified form and argue they are merely 
>INFORMATIONAL. Neither transform requires much by way of a specified 
>feature. If we eliminated the porting of a schema as a child of the 
><Transform Algorithm="&schema;"/>, all we are doing is agreeing upon the 
>algorithm URI, and repeating what the XML and schema inputs/outputs to the 
>vaidation are from their own specs.
>2. Remove the sections (but continue to leave hints that schema and XML 
>validation should be treated as transforms).
>4. Remove the sections and place them in the Auxillary Algorithms draft?
>Whatever we do, we *might* have to bounce back down to a last call or CR 
>before going to REC for a few weeks, but I'm less concerned with that then 
>getting consensus on a good decision on our options above.

Received on Friday, 21 September 2001 10:20:32 UTC