Re: [Lsid-developer] xmlns definitions at OMG ?

[Boris S. <bor...@ut...>]

Thanks for your reply, Stephen -

I take your point to mean the file locations are not guaranteed to be 
valid (or am I misunderstanding you and what you mean is that they 
*must* exist but it is up to the client's XML processor to discover 
where the really are?).

Whatever, as a service *provider* I guess I should take up the 
responsibility to ensure all URIs are valid - otherwise the model 
cannot be made semantically consistent. I am not aware of a discovery 
mechanism for xmlns.

So the Right Thing would appear to be: mirror all the definitions, 
point to my mirrored version, run a script to check if any of them get 
updated, and if they do, check whether this has implications for my 
documents, then update definitions and code as needed.

Now, while that's oK and doesn't appear too much of a hassle, I'm 
pretty sure *nobody* actually does it this way. Otherwise missing 
definitions would raise alarms. Does this mean the real world does not 
need them anyway?

Or what does the real world out there do?
Warm regards


Boris

====================================

On Monday, Jun 7, 2004, at 11:33 Canada/Eastern, Stephen A Evanchik 
wrote:

>
> Hi Boris,
>
> The locations are URIs and as such may or may not be at that location. 
> It is left up to the XML
> processor to put those namespace and location URIs in to a meaningful 
> context.
>
> You can find the WSDL files in the archive linked from this page:
>
> http://www.omg.org/cgi-bin/doc?dtc/04-05-02
>
> I am also going to include them in the next Perl release, but you 
> should always check with the OMG for the
> most up to date versions.
>
> Hope that helps.
>
> Stephen
>
> lsi...@ww... wrote on 06/04/2004 
> 03:37:23 PM:
>
> > Hi,
> >
> > The WSDL returned by the Perl LSID authority contains references to
> > definitions (e.g. xmlns:dhb) and imports - all located at
> > http://www.omg.org/LSID/2003/...  None of these files seem to exist.
> > Does this not formally invalidate the WSDL document, since it 
> actually
> > contains tags using these definitions?
> >
> > I'm trying to understand whether (as a data/service provider) I would
> > need to mirror such files on my own machines and refer to these, 
> rather
> > than definitions elsewhere, in order to be able to guarantee that the
> > URLs actually exist?
> >
> > Thanks for some enlightenment.
>