• To: Joe English <jenglish@f...>
• Subject: Re: RELAX NG Marketing (was RE: Do Names Matter?)
• From: Francis Norton <francis@r...>
• Date: Wed, 27 Mar 2002 14:09:47 +0000
• Cc: xml-dev@l...
• References: <682346532.1017169818@[192.168.0.75]> <DC1124C4368AA9438D9DEC34ECB6F162013EB79E@o...> <200203261528.g2QFSQs21121@d...>
• User-agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:0.9.9) Gecko/20020311

Joe English wrote:

>I don't think this is nearly as important as people perceive
>it to be.  XML vocabularies that are defined by a RELAX NG
>schema do *not* require applications to include a RELAX NG validator
>at processing time, unless validation is specifically required as
>part of the processing chain.
>
>Using W3C XML Schemas has a much higher potential for causing
>interoperability problems: if the semantics of a vocabulary
>depends critically on the XSD-augmented Infoset, then XSD
>validation becomes a requirement at runtime, and developers
>are locked in to tools that support XSD.
>
There is a common development pattern of  developing with full message 
validation and then switching off the validation on messages from 
"reliable" sources. Can this be done with XSD, for instance if there are 
defaults specified in the schema? If not, what are the performance 
implications - has anyone measured the cost of XSD validation v. simple 
parsing?

Francis.

• References:
  • Re: RELAX NG Marketing (was RE: Do NamesMatter?)
    • From: James Clark <jjc@j...>
  • RELAX NG Marketing (was RE: Do Names Matter?)
    • From: Matthew Gertner <matthew.gertner@s...>
  • Re: RELAX NG Marketing (was RE: Do Names Matter?)
    • From: Joe English <jenglish@f...>

• Prev by Date: Re: Compiled XML
• Next by Date: Re: Capitalism and XML (was RELAX NG Marketing)
• Previous by thread: Re: RELAX NG Marketing (was RE: Do Names Matter?)
• Next by thread: RE: RELAX NG Marketing (was RE: Do Names Matter?)
• Index(es):
  • Date
  • Thread