Re: RDF query and Rules - my two cents

On Thursday, Nov 20, 2003, at 12:15 Europe/Helsinki, ext Danny Ayers 

> I haven't fully digested your arguments yet, but would be grateful if 
> you
> could answer one quick question: can't the functionality of MGET be 
> achieved
> just using mimetypes?

No. Tried that. It doesn't work. Because you can't differentiate
between request for a description of a resource, versus a request
for a representation of a resource that happens to be a description
of another resource.

Or if the resource in question has an RDF/XML representation, you
can't use content negotiation to ask for a description in RDF/XML
because how then do you differentiate between the RDF/XML representation
and the RDF/XML description.

I even tried defining a distinct MIME type for concise bounded
descriptions, but as I indicate above, in the case where you want
a description of a description, it doesn't work

In short, there are certain "meta" distinctions which the SW needs
that simply must be kept distinct from the existing web semantics.
The distinction between requests involving a representation (web)
and requests involving a description (sw) is one such distinction.

And IMO the cleanest way to implement that is with distinct methods
such as MGET, MPUT, and MDELETE.

> ...
> (a side issue that only just occurred to me was that there's nothing 
> to stop
> the results of an MGET being 500MB, but I'll save that for later ;-)

No, there isn't. But one would expect that general web solutions for
limiting the size of responses would be employed, whether the request
was GET or MGET.

Still, good point, and something to think about.



> Cheers,
> Danny.
>> -----Original Message-----
>> From:
>> []On Behalf Of Patrick Stickler
>> Sent: 20 November 2003 09:29
>> To: ext
>> Cc: ext Danny Ayers; Graham Klyne; Jim Hendler; Dan Brickley;
>> Subject: Re: RDF query and Rules - my two cents
>>> To allow the SW to function and scale as efficiently as the Web, 
>>> there
>>> needs to be the same degree of transparency in the requests for
>>> descriptions
>>> as there is in the requests for representations. For the web, all you
>>> need to do to get a representation is use GET with a URI that is
>>> meaningful
>>> to the HTTP protocol. For the SW, all you need to do to get a
>>> description
>>> is to use MGET with a URI that is meaningful to the HTTP protocol.
>> I.e., to GET a representation, you don't GET a disk image, from which
>> you can extract the single representation desired. Likewise, for the 
>> SW,
>> you wouldn't GET an entire model/KB/graph from which you would extract
>> the resource description desired.
>> GET is resource-centric, for representations. MGET is also
>> resource-centric,
>> for descriptions.
>> If folks want to GET entire models, fine, more power to them. But I
>> think
>> that most SW agents will be far more interested in obtaining knowledge
>> about particular resources (terms, events, persons, servers, 
>> documents,
>> etc.) and won't want (nor should have) to bother with models, 
>> databases,
>> files, etc.
>> The URI alone should be enough for the most fundamental form of SW
>> activity,
>> just as it is for the Web.
>> Cheers,
>> Patrick
>> --
>> Patrick Stickler
>> Nokia, Finland

Received on Thursday, 20 November 2003 05:38:29 UTC