Re: SKOS use cases format

Hi Alistair,

Having read the other use case documents Guus pointed to [1], I would 
say that your proposal offers quite a nice setting. Lots of stuff there 
is indeed organized according to a general pattern including general 
introduction, scenario depiction, example, and problems in terms of 
standard being built.
So something quite similar to your frame, and I think we could start - 
as you did for the SWED case - to gather contributions using it. But of 
course there are comments/questions I would like the WG to discuss (even 
if to discard them quickly)

1. Level of detail
Your format seems to lead to quite elaborate descriptions, perhaps more 
than the ones in the mentioned UCR docs. I think this is not a problem, 
but I would like to have advice from people experts in that kind of 
technical document!


2. Independance of vocabulary section with respect to functionality section
I think that from our SKOS perspective it's important to emphasize on 
the vocabulary section for use case description. Even if you make the 
point in [3] that application focus is crucial, SKOS is finally about 
representing vocabularies. And I believe it's important for use case 
providers that they can express their needs with respect to this core 
aspect of their business. And therefore to do it in a section thay can 
immediately identify.

This makes me ask again the question of a more fundamental distinction 
between functional requirements and representational ones. A same 
functional scenario (described in the "functionality" section) might 
apply to several vocabulary scenarios, and vice versa: a same vocabulary 
can be used in several types of application.
I think this might confuse use case contributors: one could recognize in 
an existing use case a general application scenario he wants to specify, 
but not the vocabulary his company has to use. In such a situation, 
should the contributor "copy-paste" the functionality section? Or should 
he try to find his own word for his own scenario, even if something 
similar exists?

A possible solution could be the division of use cases in two lists of:
- "functional" cases, describing systems that exhibit specific behaviours,
and
- "representational" cases, presenting specific vocabularies with 
features SKOS should (or should not) enable one to represent

But I'm not really sure that this is a workable solution, as the 
motivation for both aspects comes from the fact that they occur in a 
same application. So formal links between the 2 lists would be needed, 
at least. And perhaps it's the work of the editor to let the contributor 
fill in a complete use case document, and the work of the editors to 
recognize the identical contributions and deal with it. What do you 
think of that?

3. Link to ISO standards.
Guus mentioned in [4] that we should link the use case to ISO standards. 
I think we should encourage the contributors to do so, if their case is 
already linked to it. I favor the addition of a "(non)compliance with 
existing encoding/representational standards" item in the vocabulary 
section. But I think we should mention the fact that filling this item 
is not mandatory, some vocabularies being developped outside of such 
considerations. Notice that amongst the existing standards, we could 
also mention the current version of SKOS! If a contributor already has 
insight on that this will provide with ammunition for the requirement 
list or the issues list that we've got to maintain.


Best,

Antoine

[1] http://lists.w3.org/Archives/Public/public-swd-wg/2006Nov/0028.html

[2] http://lists.w3.org/Archives/Public/public-swd-wg/2006Nov/0014.html

[3] http://lists.w3.org/Archives/Public/public-swd-wg/2006Nov/0023.html

[4] http://lists.w3.org/Archives/Public/public-swd-wg/2006Nov/0032.html

Received on Monday, 20 November 2006 09:51:53 UTC