[Zope-Moz] Re[2]: [Zope-Moz] RDF Use cases

Martijn Pieters Martijn Pieters <mj@digicool.com>
Mon, 20 Dec 1999 17:38:30 +0100


Hello Brian,

I wrote:
>> The above description of the use of RDF in Zope is but one example,
>> albeit an extensive and far reaching one. Zope would not actually be
>> "doing" any RDF, it would only generate XML to represent a RDF
>> datamodel.

And on Monday, December 20, 1999, 5:35:24 PM, you wrote:
BL> I would disagree with that - at least at this point. While we may
BL> decide in the end that supporting abstract RDF data model handling
BL> in Zope may be more than we want to tackle right now, we shouldn't
BL> prejudice our requirements process by ruling it out just yet.

Oh dear, that's absolutely not what I meant. What I meant is that in
the example I described, to emphasize the difference between RDF and
it's XML description, no actual RDF was used.

BL> I could imagine, for instance, an architecture where Zope 
BL> provides a way of working with RDF Model objects, and a
BL> developer could implement a new RDF vocabulary by building 
BL> an appropriate Model object. There could be standard machinery
BL> (maybe even in the Model object) for serializing the Model to
BL> xml. Perhaps it would even be possible to do controlled 
BL> intersections and unions of Model objects so that developers 
BL> could easily build compound datasources by just joining together
BL> some existing vocabularies and then serializing it.

I think I actually said something like this in another email; being
a Dutchman, my jargon may have been somewhat lacking though... =)

>> SC> Also, why not have the Zope Studio also query and 
>> manipulate the Zope
>> SC> objects through RDF only (vs having some other interface).
>> 
>> Because we want to keep Zope as open as possible. RDF would be one way
>> of representing Zope data. WebDAV another. XML-RPC is one way of
>> manipulating Zope, SOAP another.

BL> They are not mutually exclusive though. My 2 cents on this is
BL> that right now the first priority is RDF _production_, and 
BL> that object manipulation from RDF is not within the scope of
BL> the current effort. That said, an important tenet of our 
BL> development process is that we try to design a stable 
BL> architecture at the outset that will not only solve the 
BL> immediate problem but that we also think will be able to 
BL> handle future needs. So while _implementing_ manipulation 
BL> from RDF is not within the scope of this project, it sounds
BL> like making sure that our architecture does not preclude that 
BL> possibility _is_ within the scope of the project.

Thanks Brian, this clears up a question I had for you as well. =)

-- 
Best regards,
 Martijn                            mailto:mj@digicool.com