LegalRuleML Inaugural Technical Meeting



LegalRuleML Technical Meeting

Minutes July 11th 2012

Time

LegalRuleML TC meeting – July 11th

9,30pm CET

5,30am Australia

3,30pm EDT

12,30pm PDT

Attached document:

1. minute of June 29th

Participants

|Individual |Tara Athan |Voting Member |

|RuleML, Inc. |Harold Boley |Voting Member |

|University of Bologna-CIRSFID |Giuseppe Contissa |Member |

|NICTA |Guido Governatori |Chair |

|University of Bologna-CIRSFID |Monica Palmirani |Chair |

|RuleML, Inc. |Adrian Paschke |Secretary |

|University of Liverpool |Adam Wyner |Secretary |

Agenda

* Establish a versioning system for a library of xml examples

* Namespace issue for the LegalRuleML outcomes and relationship with RuleML (proposal: to duplicate the tags)

* Deliberation of the 2.1.1defeasibility.006.doc, consolidated version including the last TC amendments

* Discussion of the Tara RDF model on isomorphism

* Presentation and discussion of the 2.8isomorphism 002 003.doc from Adrian

Monica opened the LegaRuleML TC meeting in time.

In total 23 people members. Only 9 people have the voting right. Monica checks who is present: 7 people are in the call. As there is a quorum, the meeting starts. Each participant registered the attendance into the OASIS Kavi system.

The agenda is presented for collecting further amendments or integrations.

No integrations are required by the attendees.

Minutes of the previous TC Meeting

The TC approved the minutes of the previous meeting June 29th with an amendment from Harold.

The sentence:

“After a long discussion on the name and on the opportunity to introduce this mechanism, the TC approved this main principle, apart the name of and that we will decide later.”

Is modified in:

“After a long discussion on the name and on the opportunity to introduce this mechanism, the TC approved in principle the issue, apart the name of and that we will decide later.”

It is introduced also a new best practice in skype in order to make evident the voting mechanism on a motion. We declare before the motion that we intend to put in voting using:

“PROPOSED MOTION:” linguistic locution.

After a proper an fair discussion where all the parties are able to express their opinion, the proposer changes the "PROPOSED MOTION:” in "MOTION" and call the voting explicitly.

Only after this procedure the voting of the MOTION is valid.

Discussion on URI

The discussion was focused on the URI and the possible methodology to manage them in LegalRuleML. From the discussion appears evident that there are different vision of that.

Somebody suggest to formalize the requirements in plain English other to use the XML syntax because we need to test the technical feasibility and compatibility with RuleML.

Monica and Guido: The requirement is to manage a pointer to a fragment of the rule (atom, if, then, etc.) or to a rule for different purposes:

a) to connect the fragment of the rule or the rule to specific legal metadata called “legal context information”;

b) to connect the fragment of the rule or the rule to the textual legal sources e.g. Akoma Ntoso files but not limited only to Akoma Ntoso standard. LegalRuleML is independent from any legal ontology (with some preference for LKIF-core) and from any XML legal document standard;

c) to reuse some fragment of the rule or a rule in other rulebase (recall the rule without rewrite it). All these references could refer to an internal XML node of the same physical XML file or to an external XML file or in case of the textual legal sources also to non-XML file (e.g. PDF). For this reason we need a pair of attribute (key, keyRef or node and noreRef, or id and idRef) with data type anyURI for permitting the double usage.

Adrian propose three kind of attributes like in Reaction Rule: key, iri, node with different meaning. This for taking also in consideration the computational side effect on the Xpaht manipulation.

Monica: the issue is important and we will take in consideration for sure, nevertheless the computational problem could be resolved with 10 lines of XSLT as in Akoma Ntoso.

Harold, Tara: suggested only one pair and also Guido and Monica agree.

After a long discussion on the URI and on the resolvers the TC approved that:

MOTION: LegalRuleML does not need to be 'aware' of any resolver. Instead, LegalRuleML should have mechanisms to unambiguously point to (parts of) legal documents, textual provisions or legal rules.

Discussion on establish a versioning system for a library of xml examples

The TC approved to use the SVN platform provided from OASIS.

Discussion on RuleML and LegalRuleML relationship among the two standards

LegalRuleML charter defines the goals of the TC. In particular LegalRuleML wants to define a better syntax for representing the peculiar characteristics of the legal domain in agile way, with a non verbose syntax, in line with RuleML. The intention is not only to provide a methodology how to reuse the existing RuleML Familiy standards (e.g. Reaction Rules+Delibarative Rules+etc.), but also to obtain as an outcome a separate xsd schema (also in RelaxNG).

The TC discussed also on the idea to use a namespace for clearly tagging and provenience of them (from RuleML o from the LegalRuleML TC).

PROPOSED MOTION: A deliverable of the LegalRuleML TC is a schema for XML (in Relax NG and XSD) defining LegalRuleML as an extension to RuleML in the form of a combination of sublanguages from different namespaces combining at least the RuleML namespace and the LegalRuleML namespace.

Deliberation of the 2.1.1defeasibility.006.doc, consolidated version including the last TC amendments

We will be approved the next TC.

Other business

No other business.

Adjournment

The conference finished at 5.40pm EDT and it was adjourned with the pending issues in the agenda.

Tentative Agenda:

* Namespace for LegalRuleML outcomes and relationship with RuleML (15)

* Adrian presentation on isomorphism (15min)

* Tara's RDF model on isomorphism (15min)

The next meeting will be held July 25th in Skype, 4.30pm, EDT, 10.30pm CEST, 9.30pm BST, 12,30pm PDT, 6,30am Brisbane Time

Annex from the Skype conference call July 11th

[11/07/2012 21:22:29] mp: 10 minutes to the TC meeting

[11/07/2012 21:27:01] Adam Wyner: are we having the TC on skype or on the adobe thing?

[11/07/2012 21:27:18] Guido Governatori: skype

[11/07/2012 21:27:20] mp: Adam only the LegalDocML is on Adobe

[11/07/2012 21:27:24] Adam Wyner: ok.

[11/07/2012 21:27:33] Adam Wyner: Sorry, mixing what happens where.

[11/07/2012 21:29:50] mp: Please remind to record your attendance in the OASIS Kavi. Thanks

[11/07/2012 21:31:29] mp: Adam can you take the minutes for this time?

[11/07/2012 21:31:56] Adam Wyner: ok, but may have to restart my connection....something not working.

[11/07/2012 21:32:15] mp: or Adrian in case

[11/07/2012 21:32:31] Adam Wyner: I seem to be OK with skype at the moment, but I cannot access oasis.

[11/07/2012 21:32:35] Adam Wyner: or the bbc.

[11/07/2012 21:32:40] mp:

[11/07/2012 21:33:29] mp: may I record your attendance Adam?

[11/07/2012 21:33:32] *** Conference call ***

[11/07/2012 21:33:50] Adam Wyner: yes, please record me. I have access to oasis now.

[11/07/2012 21:34:06] mp: Harold can you record your attendance in Kavi ? thanks

[11/07/2012 21:34:27] mp: Agenda

* Establish a versioning system for a library of xml examples

* Namespace issue for the LegalRuleML outcomes and relationship with RuleML (proposal: to duplicate the tags)

* Deliberation of the 2.1.1defeasibility.006.doc, consolidated version including the last TC amendments

* Discussion of the Tara RDF model on isomorphism

* Presentation and discussion of the 2.8isomorphism 002 003.doc from Adrian

[11/07/2012 21:34:56] mp: ok we can start

[11/07/2012 21:34:57] Adam Wyner: The agenda has been introduced.

[11/07/2012 21:37:13 | Edited 21:42:55] Harold Boley: "mp: MOTION: to approve this principle:"

[11/07/2012 21:37:21] mp: TC approved this main principle, apart the name of and that we will decide later.

[11/07/2012 21:42:22] Adam Wyner: Discussion about how to make and vote on proposals.

[11/07/2012 21:44:56] Harold Boley: Tara/Adam: First put intention in English, only then look into XML syntax.

[11/07/2012 21:46:26 | Edited 21:47:41] Harold Boley: The English is on a higher level so stays stable, the XML syntax may be unstable since it may be subject to variation through discussions.

[11/07/2012 21:47:46] Adam Wyner: The purpose of expressing the proposal, as Harold and Tara suggest, is also to support communication later to others about the meaning and use of the syntax.

[11/07/2012 21:50:43 | Edited 21:51:20] Adrian Paschke: we only discussed on the syntax level, but did not discuss what implications it has on being compliant to standard XML. If keyref is a unique keyref within one XML instance and at the same time an IRI pointer to another XML instance, it has implications on the XML processing of a Legal RuleML document instances.

[11/07/2012 21:51:25] Tara Athan:

[11/07/2012 21:52:07] mp:

[11/07/2012 21:53:53] Adrian Paschke: Reaction RuleML uses iri for referencing to external documents

[11/07/2012 21:55:57] Adam Wyner: Guido - legal knowledge can be distributed over several documents.

[11/07/2012 21:56:44] Adam Wyner: Guido - we need a mechanism to identify portions in the different documents to relate the distributed knowledge.

[11/07/2012 21:58:06] Adrian Paschke: distributed knowledge bases are also supported by Reaction RuleML but it uses @iri to point to it

[11/07/2012 21:58:51] Tara Athan: #rule1

[11/07/2012 21:58:54] Adam Wyner: Tara - a distributed identifier using URI.

[11/07/2012 21:59:21] mp: #rule1, uri-reference, uri

[11/07/2012 21:59:48] Adrian Paschke: but you cannot ensure that @iri is actually pointing to the right key in the other XML instance

[11/07/2012 22:03:00] Adrian Paschke: it is of course alway possible to write additional pre-processesors (using XSLT or other languages), life cycle managers, URI dereferencing APIs etc. but then we are outside of standard XML processing. We then need to provide a set of tools for Legal RuleML

[11/07/2012 22:04:10] Adam Wyner: Tara - could you please summarise the issues and points as you understand them so we can record it in the chat....?

[11/07/2012 22:06:29] Harold Boley: The TC Charter says ():

"The goal of the LegalRuleML TC is to extend RuleML [RuleML 2011] with

features specific to the formalisation of norms, guidelines, and legal reasoning."

[11/07/2012 22:07:53] Tara Athan: There seem to be 3 levels of identification- within document (such as #rule1), a general external reference that is a URI, and one that is not familiar to some of us - a non-URI identifier (Akoma Ntoso, Lexis-Nexis, etc) that follows its own systematics. We will need to refer to all such identifiers from LegalRuleML, possibly multiple different systems of non-URI identifiers within the same document.

[11/07/2012 22:09:12] Adam Wyner: Thanks Tara. Question - how does this address the issues that Monica and Guido raised about distributed legal knowledge. You seemed to have a view on how this could be done.

[11/07/2012 22:10:16] Tara Athan: Adam: it is quite complex. I have some ideas about how I could meet these needs in the sytax, but I have not worked out the details.

[11/07/2012 22:10:54] Adrian Paschke: Reaction RuleML has support for distributed knowlede bases. For instance, here an example from the very old Reaction RuleML 0.2.

r1

...

...

r2...

[11/07/2012 22:11:19] Adam Wyner: OK. Thanks. Is it possible to have a document with a URI, where that document refers to the distributed information? In effect, the 'collector' document would hold references to the distributed knowledge....? Does that make any intuitive sense?

[11/07/2012 22:11:38] Adrian Paschke: In RuleML 1.0 and Reaction RuleML 1.0 oid has been replaced by @node and @key to distinguish the ambigious semantics of

[11/07/2012 22:12:05] Adrian Paschke: and Reaction RuleML 1.0 uses @iri to point to external document (instances) and keyref to point to a unique key within one document

[11/07/2012 22:12:33] mp: adam can you take track on this discussion?

[11/07/2012 22:12:47] Adam Wyner: Adrian, I don't read the reaction rule proposal as quite the same thing as what we are discussing.

[11/07/2012 22:13:02] Adam Wyner: Monica - I'm trying to track the discuss, yes.

[11/07/2012 22:13:38] Adrian Paschke: my comments are still about the node - noderef issue which we started from

[11/07/2012 22:14:27] Adam Wyner: The difference is that one is having rules with some further specification as opposed to having simultaneous information holding distributively.

[11/07/2012 22:16:08] Adam Wyner: Some discussion about differences of RuleML and LegalRuleML needs.

[11/07/2012 22:16:51] mp: Discussion on the need to provide clear outcomes of LegalRuleML to provide in front of the OASIS board

[11/07/2012 22:16:55] Adam Wyner: Main differences between RuleML and the needs of LegalRuleML - defeasibility and deontics.

[11/07/2012 22:17:04] mp: temporal legal reasoning

[11/07/2012 22:17:44] mp: legal metadata

[11/07/2012 22:21:09] Adam Wyner: There appear to be differences between what can or cannot be covered by RuleML generically (legal metadata? temporal legal reasoning?) and not covered (defeasible rules and deontics?) Have we clearly demonstrated what RuleML can/can't do to clarify to everyone why change/modifications/additions are needed? There appears to be some questions about this point. Harold seems to be proposing that various extensions to RuleML might be useful for LegalRuleML. Monica is concerned with what can be defined within the LegalRuleML TC.

[11/07/2012 22:23:52 | Edited 22:24:05] Harold Boley: See diagrams in

[11/07/2012 22:28:06] mp: ---- LegalRuleML ----

/ / | \ \

------------------- / | \ LegalRuleML Metadata

/ / | \ \ \ \

Deontic / | \ sources authors normsQualification

LegalRuleML Defeasible RuleML Temporal \ \

\ | \ LegalRuleML \ jurisdiction

\ | LegalRuleML/ \ \

\ | argumentation \ \

\ | / \ \

Deliberation RuleML Reaction RuleML

[11/07/2012 22:29:00] mp: LegalRuleML is independent from any legal ontology, any legal XML markup standard for the textual provision

[11/07/2012 22:30:05] mp: uri-reference

[11/07/2012 22:31:02] Adrian Paschke: have not known about this resolver before? When was this discussed in the TC?

[11/07/2012 22:31:06 | Edited 22:33:43] Adrian Paschke: so we would introduce another additional tool which is required to use before we can Legal RuleML XML document?

[11/07/2012 22:31:37 | Edited 22:31:44] mp: /au/act/2012/main.xml

[11/07/2012 22:32:16 | Edited 22:33:11] Adrian Paschke: is it compliant to other XML technologies, e.g. JAXB which is used by many rule engines to translate from the XML rule interchange language to the platform specific language?

[11/07/2012 22:32:41] mp: [mercoledì 11 luglio 2012 22:31] mp:

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download