Re: says it's valid, but it isn't even well-formed

... so it seems we have to agree to disagree. But building the new 
validator on Xerces-C will satisfy most of my needs, so I'm happy enough :)

Tobi

Terje Bless wrote:

> Tobias Reif <tobiasreif@pinkjuice.com> wrote:
> 
> 
>>Terje Bless wrote:
>>
>>>The limitations in OpenSP's XML support are:
>>>
>>I'm looking forward to when a real XML validator will be used.
>>
> 
> OpenSP *is* a "real XML validator"; more so, in fact, then most as it
> actually treats XML as a limited profile of SGML. :-)
> 
> 
> 
>>It would be best if several choices would be offered to the user:
>>
>>  * Xerces-C (eg DOMCount or other Xerces-based)
>>  * Xerces-J
>>  * xmllint
>>  * others
>>
> 
> I disagree. The inner workings of the Validator should be kept from the
> user when it is possible to do so. Offering multiple backend parsers to end
> users carries a cost in usability and maintenance/support with little
> actual value gained. Having support for multiple back-end parsers is good
> from a development perspective -- and is allready on our TODO list for that
> reason -- but probably with only one officially supported backend at any
> given time.
> 
> Right now it looks like that backend will be Xerces-C[0], with possibly an
> option for OpenSP since we need that for SGML in any case.
> 
> 
> 
> [0] - Well, Xerces-P, actually, but as these are just Perl bindings for
>       Xerces-C it amounts to mostly the same thing.
> 
> 
> 


-- 
http://www.pinkjuice.com/

Received on Friday, 13 December 2002 04:38:45 UTC