Re: WSDL Import/Include Locations

I wasn't aware that schema getting it wrong meant that we also had to 
get it wrong. Networks are unreliable and having a single location to 
retrieve critical data is a single point of failure. Single points of 
failure in unreliable networks seems a bad thing. So let us not do it.

Martin Gudgin wrote:

> I'm not sure why you'd draw that conclusion. I note schema import has
> exactly the same issues as wsdl:import and I've not heard anybody
> suggesting that schema is broken in this regard...
> 
> Gudge
> 
>  > -----Original Message-----
>  > From: www-ws-desc-request@w3.org
>  > [mailto:www-ws-desc-request@w3.org] On Behalf Of Yaron Y. Goland
>  > Sent: 01 March 2004 23:02
>  > To: Martin Gudgin
>  > Cc: Arthur Ryman; www-ws-desc@w3.org
>  > Subject: Re: WSDL Import/Include Locations
>  >
>  >
>  > Which means that if one did decide to use two imports to pull
>  > in the same data for the purpose of robustness then if both
>  > imports succeed, besides wasting time and resources, the WSDL
>  > would be illegal since the same data would be pulled in twice.
>  >
>  > Therefore it would seem appropriate to make it possible to
>  > list multiple URIs on a single import with the explicit
>  > semantics that one should try them in a random order (for
>  > load balancing purposes) until one works.
>  >
>  >       Yaron
>  >
>  > Martin Gudgin wrote:
>  >
>  > > I didn't say the spec prohibited two or more imports with the same
>  > > namespace attribute. I said it prohibited two or more imports that
>  > > resulted in components in a given symbol space with duplicate names.
>  > > 
>  > > I believe the spec already states that an import is a necessary
>  > > condition in order to reference components in a given
>  > namespace that
>  > > is not the target namespace
>  > > 
>  > > Gudge
>  > >
>  > >    
>  > --------------------------------------------------------------
>  > ------------------
>  > >     *From:* www-ws-desc-request@w3.org
>  > [mailto:www-ws-desc-request@w3.org] *On
>  > >     Behalf Of *Arthur Ryman
>  > >     *Sent:* 20 February 2004 20:24
>  > >     *To:* www-ws-desc@w3.org
>  > >     *Subject:* Re: WSDL Import/Include Locations
>  > >
>  > >
>  > >     Gudge,
>  > >
>  > >     I looked at the spec and the schema and I couldn't see
>  > where we prohit two
>  > >     or more <import>'s that have the same namespace attribute.
>  > >
>  > >     Also, in the case of diamond inheritance, you will end
>  > up indirectly
>  > >     importing the same WSDL into the component model twice.
>  > We allowed this by
>  > >     saying if a component is defined twice or more then the
>  > definitions must be
>  > >     equivalent.
>  > >
>  > >     I have been looking at imports and includes in the spec
>  > and I think we need
>  > >     to clarilfy the situation. If a WSDL directly
>  > references a component from
>  > >     another namespace, then that namespace MUST be declared
>  > in a top level
>  > >     import, whereas any components that are directly or
>  > indirectly included or
>  > >     imported do become part of the component model. The
>  > component model
>  > >     "forgets" where they come from, yet we do care about
>  > where they come from by
>  > >     requiring top-level imports. We seem to have two levels
>  > at which we discuss
>  > >     correctness, namely at the document level (where we
>  > have imports and
>  > >     includes) and at the component model level (after
>  > imports and includes are
>  > >     resolved). This is analogous to PSVI. We seem to have
>  > the notion of a Pre
>  > >     and Post Import/Include Processing Component Model.
>  > >
>  > >     Arthur Ryman,
>  > >     Rational Desktop Tools Development
>  > >
>  > >     phone: 905-413-3077, TL 969-3077
>  > >     assistant: 905-413-2411, TL 969-2411
>  > >     fax: 905-413-4920, TL 969-4920
>  > >     intranet: http://w3.torolab.ibm.com/DEAB/
>  > >
>  > >     www-ws-desc-request@w3.org wrote on 02/18/2004 06:05:01 PM:
>  > >
>  > >      >
>  > >      >
>  > >      > Martin Gudgin wrote:
>  > >      > Actually I think our current spec already prohibits
>  > multiple imports
>  > >      > ( or includes ) of components with duplicate names.
>  > So while you
>  > >      > could do two imports of the same namespace, it would
>  > only work if
>  > >      > all the components in the second had names different
>  > from those in the
>  > >     first.
>  > >      > 
>  > >      > Gudge
>  >
>  >
>

Received on Wednesday, 3 March 2004 18:28:41 UTC