RE: NEW ISSUE: Schema tweaks

One more issue we should consider:

Our schema defines RetryAfter as an xs:NonNegativeInteger.  The lexical
space of this data type allows an arbitrary number of digits.  It might
simplify implementations and possibly help interoperability to have a
defined upper bound for this value.  There is a type xs:unsignedLong
which is derived from xs:NonNegativeInteger, with a maxInclusive value
of 18446744073709551615.  That certainly seems sufficient for any
reasonable use, and in some cases may be more amenable to processing by
tooling.  I propose we define RetryAfter as xs:unsignedLong.

> -----Original Message-----
> From: David Orchard [mailto:dorchard@bea.com]
> Sent: Sunday, February 27, 2005 6:49 AM
> To: Jonathan Marsh; public-ws-addressing@w3.org
> Subject: RE: NEW ISSUE: Schema tweaks
> 
> +1 to #3
> 
> On rationale for using blockDefault, Tim Ewald has some good comments
> at
> http://pluralsight.com/blogs/tewald/archive/2004/08/24/2020.aspx
> 
> blockDefault="#all" basically says substitution groups at runtime are
> not allowed.  Imagine that somebody made their own type and said it
> substituted for an EPR and then plunked it into a ReplyTo, yuck.
> 
> Cheers,
> Dave
> 
> > -----Original Message-----
> > From: public-ws-addressing-request@w3.org
> [mailto:public-ws-addressing-
> > request@w3.org] On Behalf Of Jonathan Marsh
> > Sent: Sunday, February 27, 2005 6:24 AM
> > To: public-ws-addressing@w3.org
> > Subject: NEW ISSUE: Schema tweaks
> >
> >
> > In looking at the schema at
> > http://dev.w3.org/cvsweb/~checkout~/2004/ws/addressing/ws-addr.xsd
> and
> > see a few improvements that could be made.
> >
> > Title: Suggested schema tweaks
> >
> > Description: There are several ways to improve the schema, as
> proposed
> > below.
> >
> > Justification: Improve the schema :-)
> >
> > Target: Schema
> >
> > Proposal:
> >
> > 1) Add attribute wildcards to ReferenceParamatersType and
> PoliciesType
> > so there is consistent attribute extensibility throughout.
> >
> > 2) The spec says @RelationshipType defaults to
> > "http://www.w3.org/2005/02/addressing/reply" but this isn't
> indicated
> in
> > the schema.  Although we don't want to require schema validation in
> > order to construct the property value correctly, is there a reason
> we
> > shouldn't describe this defaulting in the schema?
> >
> > 3) Add @blockDefault="#all" to xs:schema.
> >
> > 4) Add @elementFormDefault="qualified" to xs:schema.
> >
> > 5) We seem to mix styles on the names of types.  Some end in "Type",
> > others don't.  Suggest adding "Type" AttributedURI, AttributedQName,
> > FaultCodesOpenEnum, FaultCodes, AttributedNonNegativeInteger.
> >
> >

Received on Monday, 28 February 2005 19:10:31 UTC