Re: [xml-dev] What is Data?

[Frank Manola <fmanola@acm.org>]

On Sep 2, 2009, at 11:03 AM, Peter Hunsberger wrote:

> Went off list, by mistake.  Uche comments on RDF  got me to ask what
> he meant and it seems relevant to the general question....
>
>
> ---------- Forwarded message ----------
> From: Peter Hunsberger <peter.hunsberger@gmail.com>
> Date: Tue, Sep 1, 2009 at 8:45 PM
> Subject: Re: [xml-dev] What is Data?
> To: Uche Ogbuji <uche@ogbuji.net>
>
>
> On Tue, Sep 1, 2009 at 6:37 PM, Uche Ogbuji<uche@ogbuji.net> wrote:
>>
>> In my opinion RDF should have been about expressing context across  
>> the
>> narrative aspects of content, including XML documents (or at least  
>> XML
>> without the mess that is  PSVI and all that).  Instead RDF makes  
>> mixed
>> content and such a pain, and focuses too much on granular data  
>> typing, and
>> relies on a basic statement model (triples) far too limited to  
>> express
>> useful nuance.

I'm still not sure what some of this means.  I could read my own  
interpretation into these things, but specifically, could we have some  
examples of what is meant by:

"expressing context across the narrative aspects of content"

"mixed content"

"granular data typing"

"useful nuance"

>
> The last part (triples) was my first frustration and I've definitely
> run into the pain of fitting mixed content into it.  Hadn't thought
> about the granular types that much, but you're right on that one too.
>  Shame, really.  Wouldn't it be nice to be able to answer Rogers
> original question by simply pointing at RDF and saying "anything that
> fits in there" with the fit, for well, everything, being obvious....
>

A variant on this might be a good way to tease out some of the issues  
people have with RDF (or with definitions of data in general).   
Suppose we answered Roger's original question by saying that data is  
anything that can fit into a relational DBMS.  The fit isn't always  
obvious (hence issues of database design), but take that as a starting  
point.  Now lets discuss the problems folks have with that  
definition.  Since anything you can fit into a relational DBMS you can  
fit into RDF (with its own issues of "database design"), presumably  
many of the problems will be the same.

--Frank

> --
> Peter Hunsberger
>