Re: [office-metadata] Binding proposal

[Michael Brauer - Sun Germany - ham02 - Hamburg <Michael.Brauer@Sun.COM>]

Hi Bruce,

Bruce D'Arcus wrote:
> 
> On Mar 21, 2007, at 3:54 AM, Michael Brauer - Sun Germany - ham02 - 
> Hamburg wrote:
> 
>> Having that said: The only choice we in my opinion have is to have ids 
>> in all documents, where most of them are unmaintained and therefore 
>> most likely not meaningful,
> 
> I think this is a little strong, in the sense that in the common case 
> (user creates new files, and saves it), the ID can be easily created 
> without their knowledge and without any issues of maintenance. The 
> common case is no problem.
> 
> The only problem areas we have identified are situations where users 
> either a) do a copy and then edit those files independently, adding 
> metadata statements to each of those changed sections, and b) creates a 
> "new" document that is not really new, but based on an existing document.

I think it is a little bit difficult to say what is the common case and 
what is not the common case, in particular since a "copy" is not 
restricted to an explicit copy by the author. A copy is also created if 
documents are attached to mails, or if someone just takes an existing 
document, maybe without the knowledge of the author,

So, the only safe area to me actually seems to be the area where 
documents are only accessed with metadata-aware office applications, and 
even not copied by the user using the file explorer or a shell. And even 
there we may not be on the safe side for the reasons I've mentioned in 
my previous mail.


> 
>> or to have ids only in those documents where they most likely are 
>> maintained only, and to have no id (or the location of the document 
>> only) for all other document. My preference actually is the later, 
>> because it allows you to differ between the document with (most 
>> likely) maintained ids and those with unmaintained ids.
> 
> For sake of argument, if we require a document URI, it could be empty. 
> And requiring a UUID could be with the caveat that implementors should 
> consider the above problem cases and try to account for them (unless we 
> can come up with some clever solution to them now).

It seems to me that our positions aren't really far away from each 
other. It seems to me that you would like to make the id mandatory, 
while I would like to encourage implementors to provide these things, 
but would like to let the users and implementors decide whether they 
take it or leave it.

> 
> I want you to consider things from the other side, though:
> 
> If documents don't have a URI, how do you propose to allow statements like:
> 
> <http://ex.net/this> ex:draft <http://ex.net/that> .

I assume "this" is the current document, and "that" some other, and the 
statement occurs within "this".

In this case I would replace "<http://ex.net/this>" with the IRI "." 
(which denotes the current document). What to do with "that" is indeed 
more difficult. If it got an id assigned (f.i. http://ex.net/that), then 
I could take that. Otherwise I would say I cannot make that statement, 
because the author of "that" did not allow me to refer to her or his 
document.

But let me ask a different question. My understanding is the the IRI 
"http://ex.net/this"; would be contained in the "that" document, but that 
it cannot be used to locate that document. If so, how do I locate the 
document? How do I know where to search for it? On my hard drive only? 
Or in the internet? What do I do if I find several documents that claim 
to be "http://ex.net/that";?

> 
> ...? Or say a series of chapters files in a book:
> 
> <http://ex.net/book/chapters/1> dcterms:isPartOf <http://ex.net/book> .
> 
> If you leave the URI optional, you force users to take on the burden of 
> identification. And if things aren't identified, they cannot be referenced.

Sorry, bit I need more details to think about this example.

> 
> Bruce
> 

Michael
-- 
Michael Brauer, Technical Architect Software Engineering
StarOffice/OpenOffice.org
Sun Microsystems GmbH             Nagelsweg 55
D-20097 Hamburg, Germany          michael.brauer@sun.com
http://sun.com/staroffice         +49 40 23646 500
http://blogs.sun.com/GullFOSS

Sitz der Gesellschaft: Sun Microsystems GmbH, Sonnenallee 1,
	   D-85551 Kirchheim-Heimstetten
Amtsgericht Muenchen: HRB 161028
Geschaeftsfuehrer: Marcel Schneider, Wolfgang Engels, Dr. Roland Boemer
Vorsitzender des Aufsichtsrates: Martin Haering