Re: proposal to have sequential / grouped messages in soap output

Oops. This had slipped past my radar. Sorry about my previous message  
being silly without taking this into account.

On Oct 18, 2007, at 03:23, olivier Thereaux wrote:

> * build a new version (from scratch if necessary, although reusing  
> the work done in the w3c validator, as well as Henri's[1], would be  
> a good idea) of the API
[...]
> [1] http://wiki.whatwg.org/wiki/Validator.nu_XML_Output
>
> Henri, would you be interested in merging the two outputs together?

That depends on what changes it would entail to the Validator.nu  
native format. I don't want to add SOAP cruft to the Validator.nu  
native XML format.

I had a look at the HTML output of the W3C Validator. It seems to me  
that the Validator.nu XML format is capable of conceptually capturing  
all the non-redundant data that is exposed in the HTML output of the  
W3C Validator except the parse tree and the W3C Validator  
implementation-specific error id numbers.

The outcome of the validation (success/failure/indeterminate) is not  
explicitly communicated by the Validator.nu XML format as the outcome  
can be unambiguously computed from the data that is explicitly  
communicated. Hence, putting the outcome in the format explicitly  
would be redundant and would require the processing model to define  
required client action in case the client received an self- 
contradicting response.

There appear to be two minor practical issues that are not conceptual  
mismatches:
  1) The W3C Validator represents quoted markup bits as plain text  
content delimited by ASCII quotes whereas the Validator.nu XML format  
expects those pieces to be marked up using the XHTML code element.  
The Validator.nu XML format does not prevent W3C Validator-style  
message strings, but it would be nice if the quoted markup bits were  
marked up sing the XHTML code element.
  2) The message elaborations in the W3C Validator are stored as  
snippets of HTML source text but the Validator.nu XML format requires  
the elaborations to be inline XHTML fragments. A conversion from HTML  
source snippets to XHTML fragments would be needed.

As for the conceptually missing parts, the representation of the  
parse tree would need to be figured out. (The parse tree idea on the  
wiki failed review on #whatwg.) Moreover, the issue of implementation- 
specific error ids would need to be addressed somehow.

Am I missing something? What additions to the format would be required?

-- 
Henri Sivonen
hsivonen@iki.fi
http://hsivonen.iki.fi/

Received on Tuesday, 30 October 2007 12:58:07 UTC