RE: [wsbpel] Using RelaxNG

[Maciej Szefler <mbs@fivesight.com>]

Ugo,
Speaking from an implementer's point of view I am inclined to think that
not having to deal with another schema language is a false economy. What
it means is that we implementers must devise and perform many more
post-parse syntax checks to augment the presently very weak XSD
validation. The list of these checks is not maintained by the TC, so the
chances of omission, or of diversity of interpretation between
implementers is greatly increased. 

As for users, I really don't see why they'd care how the syntax of the
language is specified. Most users of most languages never bother to sit
down and read the language grammar: instead, they tend to rely on
second-hand sources, narrative text, examples, tooling, inference, and
experimentation. 

-maciej

On Mon, 2004-04-19 at 14:38, Ugo Corda wrote:
> Maciej,
> 
> So, I suspect the reaction from most BPEL implementers and users to the
> idea of using RelaxNG for BPEL language description would be something
> like: "Thank you, but no thanks. XSD already keeps me busy enough ...".
> 
> Ugo
> 
> > -----Original Message-----
> > From: Maciej Szefler [mailto:mbs@fivesight.com] 
> > Sent: Monday, April 19, 2004 12:32 PM
> > To: Ugo Corda
> > Cc: Anderson Jonathan; Ron Ten-Hove; wsbpel@lists.oasis-open.org
> > Subject: RE: [wsbpel] Using RelaxNG
> > 
> > 
> > Ugo,
> > 
> > On Mon, 2004-04-19 at 12:48, Ugo Corda wrote:
> > > Maciej,
> > > 
> > > Wouldn't that imply that a BPEL implementation would have 
> > to support 
> > > both an XSD validator AND a RelaxNG validator?
> > > 
> > > Ugo
> > Absolutely. XSD validation would still be needed for what in 
> > BPEL passes as type checking (both at run-time and in 
> > static-analysis). RelaxNG would be needed at "compile time" 
> > to check the syntax of the language. This is in line with the 
> > common decomposition of a language specification into lexical 
> > grammar (roughly ~ XML), syntactic grammar (~ 
> > RelaxNG/XSD+post-parse checking), type system (~ XSD+rules), 
> > and execution semantics(~ shared sub-conscious understanding). 
> > 
> > Perhaps others disagree, but I do not think it is remotely 
> > practical or desirable to change the typing scheme 
> > considering the close association between WSDL and XSD (not 
> > to mention the reality on the ground). 
> > 
> > -maciej
> > 
> > 
> 
> To unsubscribe from this mailing list (and be removed from the roster of the OASIS TC), go to http://www.oasis-open.org/apps/org/workgroup/wsbpel/members/leave_workgroup.php.
>