Use Case for W3C RIF-WG Wiki

> If you're not on the WG, the place to send comments and feedback is
>    public-rif-comments@w3.org
> for which you do not have to be subscribed.
> Your use cases are welcome there, or you may want to wait until after
> some discussion in the next few weeks in which the format and
> structure of use cases will be more settled.   Or you may want to wait
> and see if some other ones (like the one I drafted for the charter)
> make it into the first public working draft of the Use Cases document,
> and then offer your comments.   The Working Group is not obligated to
> accept use cases or requirements from outside, but it must be
> responsive to comments about its drafts.

If I understand it correctly, I have to send my comments and use-cases to
this list.

So, first I resend (extended) extracts from my october 23, 2005 comments

1.1 Usage scenarios.
The example "Validating Prescriptions " is confusing.
- In my opinion, for illustrating a kind of application in health it should
be modified to take into account the old tradition of thesauri and
ontologies in life sciences and also the current trend of migrating them
to OWL.
Asserting that "his pharmacy checks all prescriptions against a merged,
multi-source rule base. This rule base includes the fact that erythromycin
is a macrolide antibiotic"
(i) is not realistic:
it ignores the size and complexity of existing knowledge in the biomedical
domain area and makes the strong assumption of converting it in a *rule
base*:
(ii) is not clear enough:
In many cases, for example in bioinformatics, a main issue is the
integration of heterogeneous sources, hence the availability of a rule
language for "mapping rules". Other cases are concerned by
"merging rules". For example, merging rules are required to connect Anatomy,
together with Gene and Disease reference ontologies, which constitute the
backbone of the future Semantic Web for Life,
In that scenario, it is not clear,  for what part of knowledge a
rule language is supposed  to be used.
- For example, relating medicines to their contraindications seem to express
ordinary associations, that does not require rules.
- What "a merged multi-source rule base" means ?
does it means that rules are used only for mapping different sources to a
global ontology of drugs in OWL, or does it mean that all the
knowledge is supposed to be merged within a single repository using where
all the knowledge is represented in a rule language ?

1.2. Compatibility
In my opinion a stronger statement regarding compatibility with OWL data and
ontologies is required.
I am afraid that the use case "Product Compatibility" using a
narrative form and *instantiated with the product being a medicine*"
posted to the Wiki does not improve the case and is not really more clear.
Although, section 6. Commentary  states "This is a kind of an OWL++ use
case." and Section 3 asserts "Requires that rule bases use a shared (or
mergeable) ontology and can then be merged themselves", these points should
be clarified and revised accordingly in the scenario. It should be made
completely explicit what is represented in OWL and what is represented by
rules (see above)

I do not really understand the reasons of the change  to "Product
Compatibility", which may bring more confusion.

You may find at http://www.med.univ-rennes1.fr/~cgolb/RIF/Wiki2.pdf
another possible version for a "Validating Prescriptions" scenario
corresponding to
one of the  [Use case] from medicine  proposed by Uli Sattler at the first
f2f on my behalf , also available at
http://www.med.univ-rennes1.fr/~cgolb/RIF/OntologyWithRule-CG.pdf

Several of these use-cases are now available in the Wiki format at:
http://www.med.univ-rennes1.fr/~cgolb/RIF/Wiki1.pdf
http://www.med.univ-rennes1.fr/~cgolb/RIF/Wiki2.pdf
http://www.med.univ-rennes1.fr/~cgolb/RIF/Wiki3.pdf

Hope it helps.

As there is no much traffic on this list, I would be happy to have any
feedback.

thanks

Christine

Received on Saturday, 24 December 2005 12:51:38 UTC