Re: a different approach to OWL/XML (was: draft responses for LC comment FH3/29)

On 9 Mar 2009, at 15:46, Sandro Hawke wrote:
[snip]
> FWIW, I think procedurally we could do this.  No, it's not really our
> place to define an XML-friendly RDF syntax that everyone is  
> supposed to
> use, but if we defined one for ourselves that happened to be good  
> enough
> for everyone else to use, IMHO no one would be too upset.

Mee! mee!

Did you look at my example?

> I also think technically we could do this.  We could either select a
> subset of RDF/XML which is schema-friendly ("rigid RDF" [1])

[2]?

There is no such thing. Property serialization essentially forbids it.

> or define a
> somewhat cleaned-up version of that subset (probably not using the
> RDF/XML namespace at all, "type-tagged XML").

But this would have the same problem as Trix.

> It's something I've
> thought a lot about.
>
> Drawbacks:
>
>     - It's more work now, since OWL/XML is more-or-less done.
>
>     - OWL/XML would become more verbose; although it would probably be
>       easier for machines to parse,


What? Without question it would *harder* to parse. Also, I don't see  
how to get all the goodness of the XML tooling (e.g., type aware  
queries).

> it wouldn't be as good for the
>       hand-authoring (in XML tools) that Bijan is talking about.  I
>       suspect but don't really know that it would be about the same  
> for
>       for querying.

I'm pretty sure.

>     - OWL 2 would have two mappings to RDF Graphs, a "regular" one,
>       where each syntactic structure maps to an object with  
> properties,
>       and a more idiosyncratic on which extends RDFS (the one we
>       currently have).

What's the advantage of making a new syntax for RDF? From here? Do we  
remotely have the will or mandate for this?

> Advantages:
>
>     - We'd have a cool new bit of infrastructure, a way to transmit
>       structures so they could be understood and used by both XML and
>       RDF toolchains.

I'm pretty skeptical about the possibility of this, much less the  
feasibility.

>     - If we used rigid RDF, the GRDDL question would go away.  :-) (If
>       we used type-tagged XML, the GRDDL question would no longer be
>       OWL-specific.  I don't know if that's an advantage or
>       disadvantage.)

Oy.

> So, I think this is a good idea, and I'm willing to do work to see it
> happen, but (of course) I'll understand if the WG isn't interested or
> doesn't think it's worth it.

I'll be interested in what you produce, but I'm very hard pressed to  
see that you could accomplish something in the time/energy frame of  
the WG. And, with all due respect, I'm *really* suspicious that you  
can do something that is really and truly friendly to the XML  
Toolchain. Prima facie, this is something new! Part of what prompted  
my revision to the Schema (coming momentarily to a list near you) was  
1) using the syntax in anger for a real task and 2) boning up on the  
XML stack for my MSc class.

Cheers,
Bijan.

Received on Monday, 9 March 2009 16:04:58 UTC