RE: [plcs] Review of capabilities - Some comments from the Norwegian pil

I’m not sure if this issue has been adequately understood by all and so have a series of question I’d like to pose to the Norwegian pilot participants and to the TC in general. I have two points to make.

It seems to me that there are three possible categories of Capabilities: 1) those requiring no RD, 2) those requiring only standard RD and 3) those requiring extensions to the standard RD to be useful.

1 – Do all Capabilities depend on Reference Data? If not, have those not requiring RD been reviewed to see whether they are “specific enough” given the criteria used by the Norwegian Pilot team? If they have found to be insufficient, have specific proposals for improvements been requested?

2 – For Capabilities that do depend on Reference Data, are there any for which the standard RD is expected to be complete/sufficient for all uses? If so, have those been reviewed to see whether they are “specific enough” given the criteria used by the Norwegian Pilot team? If they have found to be insufficient, have specific proposals for improvements been requested?

3 – For Capabilities that depend on RD and for which the standard RD must be extended to be useful, there is likely to be some confusion. DEXs that use this type of Capability are not directly implementable. Therefore, an approach for handling these needs to be agreed. As an example, in the Capability or DEX itself, example RD extensions could be described in text, figures or as example OWL files to show how they are to be used.

Secondarily, I want to disagree with the idea represented by the Figure 2 in the issues document. It shows the “encansulation of one Capability within another”. I think this is a mistake. Implementors don’t want to go read a trail of 6 different capabilities to try and follow and understand an instance diagram. If the issue is knowing which entity types are described in a used capability, then some form of shading on the diagram or hyperlinking the diagram to the description in the appropriate capability is a better approach. Data exchange standards are not like OO interfaces where encapsulation is a good idea. As was mentioned earlier in the issue paper, implementers of data exchange standards are required to understand every small detail of the data structures they are populating.

Hope this helps the discussion.

Cheers,

David

-----Original Message-----
From: Trine.Hansen@dnv.com [mailto:Trine.Hansen@dnv.com]
Sent: 26 March 2004 16:31
To: plcs@lists.oasis-open.org
Subject: [plcs] Review of capabilities - Some comments from the Norwegian pilots

Finally some comments to the documentation of capabilities in DEXLib based on Norwegian pilots needs/experience.
See attached document.

<<Comments from the Norwegian Pilots after review of C032 and C016.doc>>

Regards
Trine Hansen
UCTNO940, Information Quality Management
Det Norske Veritas AS
( + 47 67 57 96 38 (office)
( + 47 90 83 44 24 (mobile)
* trine.hansen@dnv.com
http://www.dnv.com

**************************************************************
Neither the confidentiality nor the integrity of this message can be guaranteed following transmission on the Internet.
All messages sent to a DNV email addressee are swept by Sybari Antigen for the presence of computer viruses.
DNV acknowledges that SPAM email represents a potential security risk, and DNVs filters to block unwanted emails are therefore continuously adjusted. DNV has disabled "out of office" replies to the Internet
**************************************************************

plcs message