Re: [xml-dev] What is Data?

[Peter Hunsberger <peter.hunsberger@gmail.com>]

On Wed, Sep 2, 2009 at 11:31 AM, Frank Manola<fmanola@acm.org> wrote:
>
>>
>> The big difference is that RDF should be mostly all inclusive;
>> metadata (well at least partial), relationships and data all laid out
>> in one (or two if you count the schema) place.  You won't see that big
>> picture with a DBMS.
>
> You say RDF *should* be mostly all inclusive.  Should I infer that you think
> it isn't?  If so, in what way?  What are some examples?

The difficulty in putting mixed content into it is perhaps the most
glaring. However consider things like complex datatypes, sets or
arrays.  You might be able to express them in RDF but instance data
would be very painful.

> I also don't understand what you're saying about the "big picture" you don't
> see with a DBMS.  LIke what?  Is it that there's stuff you *can't* put in
> the database, or stuff that people typically *don't* but in the database?

Getting at the metadata of an RDBMs requires a separate operation from
extracting the data.

If I hand you a file representing some database (to the extent that is
possible) then, at best, it's an opaque blob until you query it's
schema (assuming you have the right tools).  From there you might be
able to make some rational queries against the actual data.

If I hand you a file that is RDF you can examine it directly (albeit,
painfully) in any text editor you might choose (assuming it can handle
the encoding).

-- 
Peter Hunsberger