xml-dev - Re: Re WF, V, and MSXML

[Michael Leventhal <michael@textscience.com>]

At 10:00 AM 6/9/97 GMT, Peter Murray-Rust wrote:
>> I would be alarmed at a processor/parser (they mean the same to me in
>> this context) that attempted to parse for validity, and if it found
>> an error, silently switched to WF-parse mode.
>
>I'd agree with this analysis, and haven't been silent on the issue.  IMO it 
>is more important for the WG/ERB to address *this* problem than some of the 
>proposed extensions.  The concept of WFness is NEW!!  It is more subtle than
>people realise.  A fundamental problem is that there is no clear internal
>flag in the document stating what the validity/WFness of the current document
>is, is meant to be, was, etc.  As Terry says, it's particularly likely that
>a WF document could (possibly erroneously) mutate into a valid one.  I am
>sure that any confusion about MSXML is not intentional and is due to the
issue
>not be prominent in the spec.  

But "silently switching" is exactly the behavior that is wanted for most
output oriented operations, e.g., browsing.  WF is only new formally but
informally it has been the default mode of operation for HTML.  I don't
think a flag stating the intention of the author could ever be supposed
to actual represent the wishes of the current user of the document or that
we could expect the majority of users to understand the underlying concept.
It is up to the user of the tool to select the mode they want if a choice
exists.

Validate and switch to well-formed "silently" is a possible mode of operation.
But I agree on requesting that each application formerly state its possible
modes of operations.

>
><PROPOSAL>
>All parsers (i.e. tools that take XML documents and apply the criteria in 
>XML-LANG only) should state their attitude and behaviour to WFness and
validity.
></PROPOSAL>
>
>The possible options include at least:
>	- nsgmls-like.  Full validation is the only option.  Any non-valid
>		dcoument is flagged and appropriate error messages or error
>		action is initiated.  
>	- Lark-like (at least V0.88 - I think there is another coming).  No
>		validation can be attempted.  Any 'output' can only be WF or
>		in error.  NOTE: what does Lark do with the internal subset?
>	- NXP-like.  Validation can be switched on or off by the 'client'.
>		How this is transmitted to the application is application
>		dependent at present.
>	- MSXML-like.  Undocumented at present.  Possibly [though Terry and I
>		hope not] validating by default, and changing to WF if this
>		fails.


xml-dev: A list for W3C XML Developers
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To unsubscribe, send to majordomo@ic.ac.uk the following message;
unsubscribe xml-dev
List coordinator, Henry Rzepa (rzepa@ic.ac.uk)