ID preservation issue WAS: Re: [topicmapmail] This is weird modeling, but is it valid XTM ?

[Kal Ahmed]

On Wed, 2003-08-13 at 21:55, Peter P. Jones wrote:
> On 13 Aug 2003 at 21:14, Kal Ahmed wrote:
> 
> > Peter,
> > 
> > In the current DM draft occurrence items and association role items
> > both have a resourceLocators property that preserves the complete URI
> > of the source element(s) (note that there can be serveral source
> > locators when duplicates are eliminated). I think that should meet the
> > need expressed by the use case you outline.
> Hi,
> 
> I'm not sure it's quite the same thing.
> By source element, do you mean the respective element of the TM 
> syntax document or of a syntax document that just happens to map to 
> the TM architectural form?

I'm not sure that I understand your distinction. If the source syntax
(what ever it is) has a mapping to a URI for the syntactic construct
that gives rise to the occurrence information item or association role
information item, then that URI should be stored on the source locators
property of the information item.

> In the latter case, it would capture providence but I would not be 
> assigning the accountability metadata to the correct thing, I don't 
> think.

What is "the correct thing" ?

> For example, if I acquire data from several sources about a single 
> subject, then assign that data to a topic that is a proxy for the 
> subject itself, and then I modify some aspect of the topic's 
> characteristic assignments, I want to be able to capture the fact 
> that I was the one that modified the topic on a certain date and 
> time.

So you want a single URI for the information item ? That is not covered
by the DM.

> It is true that if I'm altering the data then it calls into question 
> the value of the original data, but I think there are circumstances 
> where the TM becomes a knowledge resource in its own right where this 
> editing situation could arise independently of source documents in 
> other syntaxes from which the TM was originally derived.

Agreed, and you would need to have some way of doing that. For example
in TM4J, every object gets a system identifier that is not exposed as a
source locator, but is stable for the lifetime of the object. I guess
that this is the kind of thing you would be looking for. I don't think
that this is within the realms of the DM though as there is no mapping
from any syntax to that kind of identifier.

Cheers,

Kal
-- 
Kal Ahmed, Techquila
Standards-based Information Management
e: kal@techquila.com
w: www.techquila.com
p: +44 7968 529531