Re: Issue 4 Proposed Resolution (was: why no doc type declaration and PIs in SOAP)

Jacek Kopecky wrote:
> 
>  I'd just like to restate here that XML requires some DTD
> processing to be done even by non-validating parsers, so ignoring
> internal DTDs would violate XML processing rules.
>
I raised this point at the F2F when defending my original proposal[1] 
that processors MUST generate a fault on receipt of a message containing 
a DTD or PI. However it was felt that since SOAP message MUST NOT 
contain DTDs in the first place then it would be OK for a processor to 
ignore one if present. As stated in a prior message, this was to allow 
very simple processors (that presumably are without a full parser) to 
ignore DTDs and PI rather than be forced to include checks for their 
(erroneous) presence.

Regards,
Marc.

> 
> 
> On Fri, 28 Sep 2001, Marc Hadley wrote:
> 
>  > All,
>  >
>  > As the custodian of issue 4 I'd like to propose the following resolution
>  > and rationale.
>  >
>  > Proposed Resolution:
>  >
>  > A SOAP message MUST NOT contain a Document Type Declaration or
>  > Processing Instructions. On receipt of a SOAP message containing a
>  > Document Type Declaration or Processing Instruction a SOAP receiver MUST
>  > either ignore it or generate a fault (see 4.4 SOAP Fault) with faultcode
>  > of "Client.DTD" or "Client.PI" respectively.
>  >
>  > Rationale:
>  >
>  > In discussions [1,2] there is near universal antipathy towards allowing
>  > DTDs in SOAP messages. The attitude towards PIs is somewhat less
>  > negative, but is still broadly in favour of exclusion. This maintains
>  > the current status-quo inherited from SOAP 1.1.
>  >
>  > Issue 4 relates to the action a SOAP receiver should take on receipt of
>  > a message which includes a DTD or PIs. My original suggestion for
>  > resolution[1] was to require the SOAP receiver to generate a fault on
>  > receipt of such a message but this was felt to impose an unecessary
>  > burden on receivers.
>  >
>  > An alternative resolution[3] suggested relaxation of my original
>  > proposal such that receivers SHOULD ignore DTDs and PIs and MAY generate
>  > a fault but this formulation leaves open the possibility of having a
>  > compliant SOAP processor that doesn't ignore DTDs and PIs and doesn't
>  > generate a fault which I don't think is the desired behaviour.
>  >
>  > In the spirit of a friendly amendment to the preceeding suggestion I
>  > propose to give implementations the option of either ignoring DTDs and
>  > PIs or generating a fault on their receipt.
>  >
>  > Comments ?
>  >
>  > Marc.
>  >
>  > [1] http://lists.w3.org/Archives/Public/xml-dist-app/2001May/0367.html
>  > [2] http://lists.w3.org/Archives/Public/xml-dist-app/2001Sep/0159.html
>  > [3] http://lists.w3.org/Archives/Public/xml-dist-app/2001Sep/0167.html
>  >
>  >
> 
> 
> 
> 



-- 
Marc Hadley <marc.hadley@sun.com>
XML Technology Centre, Sun Microsystems.

Received on Monday, 1 October 2001 05:06:06 UTC