[RIFRAF] notes from f2f RIFRAF breakout session

Chair: Leora
Scribe: Axel

Attendees: Allen, Paula, Sandro, Hassan

Starting-points:

* Starting points for bottom-up:
    - Current list of discriminators in questionnaire
    - Attempts to fill the questionnaire
    - Attempts to ontologize sections of the questionnaire/discriminators
    - Starting points for top-down:
* REWERSE classification
    - Hassan's classification
    - The recent TED mails contains starting points
    - RuleML modularization
    - Deriving Core concepts top down from : Rule, Ruleystem, Rulelanguage,
       head, body, et.c assigning properties top-down to these from 
questionnaire.


Discussion-notes:

Sandro: Every value in the discrimintators defines a class.
Leora: But what are the slots?
Leora: We had in the phone conf beginning Oct some decision to start 
bottom-up.

ACTION: Axel to analyze top-down approaches and extract them into some 
OWL-style.

Sandro: brainstorm on top-level concepts and identify overlaps: e.g. 
condition vs. body, etc.

Sandro's approach on an OWL ontologization of Section 4 of the 
questionnaire at:
  http://www.w3.org/2006/10/rifraf/

Allen demos howto use/modify in Protegé.

Sandro: envision a system and language dialect feature-matrix.

John Hall: PRR organizes existing production rules (vendors?)
(all do productrion rules and sequential rules, i.e. ordered)

Sandro: System/Language/Implementation.

Allen: Ontology is more than a taxonomy. We should capture more than a 
feature matrix.
Define formally logical negation, strong negation, etc.?

Sandro: this is more a research question... the idea was just to capture 
ther information in the questionnaire.

Sandro: Imagine tables of dialects, these would help us to compare 
dialects with languages, how much a language can "fit" a dialect.

Sandro: rather just classes/than syntaxes.

John : presents  some of his ideas in  business rules languages, 
mentioning aggregates (count/average)

Remark Axel: This asks for whether there are aggregates in the Rule 
languages discriminators at the moment.

Leora: General reason for the document/ontology: define which languages 
to translate from/into.

Sandro:  Answers to the questionaire are quite fuzzy. We should take the 
current answers and not only the discriminators in to account!

Allen: Some discriminators seem to be lacking instances.

Sandro: Ask Harold to find an instance for each discrimintator.

Leora: New task for each of the section responsibles and check ambiguity 
of answers, whether they were unclear.

Axel: Proposal ACTION: On each of the section responsibles to do this.

Sandro:  not all systems classified in the questionnaire. We should 
should first change the questionnaire to stabilize it with the current 
answers, then go through the sys table and push people to fill at least 
these. Ontology classes/properties should have comments which translate 
maybe to a discriminator/glossary entries.

Sandro:  What people were filling out in the questionnaire is OWL 
instance data.
Can we ask people to "fill" in the next iteration, by providing instance 
data in RDF?
ACTION: Sandro to provide a template for answers, as soon first 
itertation of the questionnaire is available.

=====================================
Summarizing actions and TODO's

Results of the questionnaire so far should be taken into account:

http://www.w3.org/2002/09/wbs/38457/RAFQuestionnaire/results

We have the following Section responsibles for the parts of the current 
questionaire:
Leora: Sec 1
Hassan: Sec 2
Allen: Sec 3
Sandro: Sec 4
Paula/Leora: Sec 5
Frank: Sec 6
Proposed new parts:
   - Typing: Frank
   - Negation: Axel?

Proposed ACTION: Each of these go through filled so far answers so far
and
a) propose next iteration of features in their sections based on the 
answers.
b) extend first "ontologization" based on Sandros first attempt:
		http://www.w3.org/2006/10/rifraf

Proposed ACTION: Axel provide a first attempt for a top-down ontology
(help of the TED people appreciated! Proposed subaction.
ACTION: Harold to review this first attempt.)

Time schedule:
Sugggested: actions on all discriminators sections, and first
attempt for a top-down ontology until end of the month
First draft of a top down ontology, I will try to get that info out of 
the documents...
until end of the month.
Merge TelConf: first week ofDecember?

Document Structure:
Describe the ontology, similar to foaf: spec, describing all concepts 
and properties,
+ example languages, systems.

Procedure to collect information from the action items:

Provide a wiki-pages, RAF_v2 with subpages for each section to provide 
next iteration:
	http://www.w3.org/2005/rules/wg/wiki/RIFRAF/<secnumber>
For a)+b) part of the action above to provide at least:
- info a bout insights from answers
- ontologization

Proposed for as soon as we have a stable ontology:
Next itration: people provide instance data for their rulesystems/languages
(at least the ones in: 
http://www.w3.org/2005/rules/wg/wiki/List_of_Rule_Systems)
ACTION: Sandro to provide a template for answers as a turtle file, as 
soon first itertation of the rifraf-ontology is available.

Received on Sunday, 5 November 2006 21:37:58 UTC