Re: SWSL: Minutes for the SWSL Telecon - 21st January, 2004 #37

Dear all,

given the minutes below and the discussion topic of finite state 
machines and conversations and should we store the behavioural 
footprint it seems very sensible for you folks to track us folks (we 
obviously track you). The WS-CHOR group has taken a behaviouralist 
approach to the description of the external observable behaviour of 
participating web services. We have recently had Prof Robin Milner and 
Dr. Nobuku Yoshida and Dr. Kohei Honda join our group. These esteemed 
academics represent some of the very top people in process algebra 
(Robin invented pi-calculus). They are working with us on modeling some 
of the use cases formally so that we can understand better the 
requirements and the direction we need to go it to produce a 
Choreography Description Language (CDL). We already have a model 
overview of the CDL on our web page.

We would welcome some cross fertilization. We shall be at the plenary 
and our F2F is the Mon/Tue of that week.

Best Regards

Steve Ross-Talbot
Chief Scientist, Enigmatec Corporation
co-chair W3C Web Services Choreography
chair W3C Web Services


O: +44 207 397 8207
C: +44 7855 268 848
www.enigmatec.net


On 24 Feb 2004, at 02:54, Terry Payne wrote:

>
> [Note - these are past minutes from recent SWSL telecons]
>
> =============================================
> Minutes SWSL telecon - 21st January, 2004 #37
> =============================================
>
> Participants:
>   David Martin
>   Michael Keifer
>   Terry Payne (& Roxana Belecheanu)
>   Bijan Parsia
>   Rick Hull
>   Daniela Bernardi
>   Steve Battle
>
>
> Notetaker: TP
> Next: TP
>
> Action Items
> ------------
> 1) Steve Battle - make a proposal for a new version of the general
> requirement including objective of using OWL-S as a starting point (see
> below).
>
> 2) Look at the Service Oriented Model as proposed by Web Services
> Architecture Group; each subgroup should identify where they fit and
> provide more details.
> See http://www.w3.org/TR/2003/WD-ws-arch-20030808/
>
> Minutes
> -------
> Item 1: Date for next F2F
>
> MK: Rick Hull offered to host at Bell labs.
> RH: Can reserve hotel rooms - aprox 25 people?
> DM: Rick sent a message to Admin list proposing to host at Bell Labs,
> however details need to be confirmed asap.
>   NOTE - MK didn't receive this message
>        - check that messages are being received...
>
> MK: Any idea whether DAML-PI will be before or after?
> DM: They are thinking of holding DAML-PI after, either Mon-Wed
> or Tues-Thu
> MK: 22-23 or 23-24 would be good (the latter is better).  Is the 22nd a
> developers day???  If DAML is held on the 25-27th then that would be 
> good
>
> TP: My preference would be to hold meeting before WWW
>
> DM: any preference on ordering?
> ... F2F before WWW?
>
> DM: If DAML in Manhattan, then F2F shouldn't be sandwiched in between.
>
> Item 2: Minutes
>
> MK will post Minutes to the interest group in the coming days.
>
> Item 3: Steve's Comments on Requirements
>
> SB: Process Modelling appears to be focussed on workflow, whereas 
> DAML-S
> has focus on "conversations".  Process Model can be described as a 
> finite
> state machine.  Define a set of behavioural constraints.
> MK: Workflows can be non-deterministic etc
> SB: True - conversations are weaker...
> RH: Similar to FSA and behavioural signatures - declarative 
> specification
> (sometimes abstract) but internal details hidden).  I have a colleague 
> who
> is looking at guarded specifications - Mentor system (Gerhard's group).
> RH: The workflow describes what is behind the scenes
> RH: Behavioural Signatures - should they be in the profile?  Can also
> help with modelling and composition.
> MK: Yeah - behavioural signatures are part of the process
> SB: OWL-S describes behavioural signatures!
> RH: Mebbie we should simplify the notion of this for the novice reader
> RH: e.g. temporal constraints vs behavioural signatures, and then 
> there is
> mentor - state chart base vs work flow base
> SB: Work needs to be done to see if State Chart Base and Work Flow Base
> can share semantics or not?
> MK: Some (or all) conversations can be modeled as workflows
> SB: have sent a link
> http://www.hpl.hp.com/techreports/2003/HPL-2003-208.html
> Add to the requirements that we should cover this view.
> MK: So we add this as well as a use case
> RH: This would be a good thing!
> MK: Contact Michael G about this - send to mailing list
> SB: A second point - What is the relationship to OWL-S?
> SB: We view that OWL-S semantics should be a starting point (and 
> clarify
> semantics) - again should be explicitly stated
> MK: Have agreed to identify the gaps between requirements and OWL-S and
> use OWL-S as a starting point to reach our requirements.
> SB: Should be a non functional requirement (explicit) in Section 2 -
> general requirements.  Point 3 General requirement should be
> relabelled as "Non-Functional" requirements.
> RH: This was for general points that do not fit the other sections.
> SB: I think of general requirements as non-functional requirements.
> DM: That notion hasn't been used consistently so far...
> SB: that distinction is used in industry
> MK: I view this more as a call to action rather than a requirement
> SB: I just suggest that we should use OWL-S as a starting point.
> RH: Requirements are a reflection of all the action points :)
> would want SWSL to be an extension of OWL-S
> MK: Yes but would want to go further than this...
> MK: Would like people to volunteer to do different parts of the
> requirements.  An action item is for people to take responsibility of
> finding gaps between functional requirements and OWL-S.  E.G. to see 
> where
> it fall short of our requirements.
> SB: Needs to be a requirement to be done
> BP: There is a difference between taking OWL-S as a starting point or a
> foundation...
> SB: Where possible use the existing lang as is.
> TP: Want to push OWL-S out and standardised...
> MK: That is in the mission statement, right?
> BP: Is OWL-S a good fit for certain things for what people want?
> TP: We're not going to get a 100% solution out soon though
> MK: OWL-S is insufficient, for example in defining the full semantics 
> of
> matchmaking.
> Also look at complete examples that cannot be modelled - see example in
> F2F and last summer.  How do you represent other issues
> BP: Could you repost the example?
> MK: Soon, when mail issues are sorted.
> BP: OWL-S should consider this as a group.
> ALL: Agreed :)
>
> SB: Could what we produce be seen as OWL-S2.0 - there would be some
> lineage here...
> BP: What counts as sufficiently similar?
> SB: Can you trace the lineage back?
> TP: Will there be a departure though from OWL?
> MK: Probably to actually progress.  May stay here at syntax, but 
> modelling
> semantics will probably move on.  Different World views (DB people vs
> Logicians)
>
> BP: OWL-Rules will not support non-monotonic reasoning.
> MK: OWL does not support reasoning patterns that we might use.
> Need defaults, for example...
>
> MK: extending OWL-S should be an objective rather than a requirement
> DM: lets make this an objective in the requirements doc.  View it as a
> starting point and move on.
> MK: State this as a functional requirement in the intro.  
> Non-functional
> requirement...
> BP: But how is this a requirement?
> DM: a starting point and an objective is for future work to be based on
> OWL-S
> MK: in the introduction
> DM: agreed
> SB: in the general requirement
> BP: in the introduction
> SB: if in the into then it is not a requirement...
> BP: its only an objective anyway!!!
> MK: Steve - make a proposal for a new version of the general 
> requirement
> including this. [ACTION ITEM - Steve Battle]
>
> MK: action item to complete the conceptual model by the end of the 
> month.
> essentially an ER diagram.
> W3 Architecture group have come up with a diagram that looks relevant 
> but
> not fully detailed.
> Each of the subgroups should come up with a diagram for their
> requirements.
> Need to look at the architecture document at the SWSA site.
> MK: David mailed this out in the past couple of days. The link is in 
> the
> agenda [http://www.w3.org/TR/2003/WD-ws-arch-20030808/]
> MK: Some of the requirements should come from the arch-committee.
> Even though this model may change when the arch-committee finalise 
> their
> requirements.  W3C Arch diagram is good to identify actors - again sent
> by David.  Seems to be relevant to our conceptual model
>
> Look at the Service Oriented Model and expand [ACTION ITEM FOR ALL]
> Each subgroup should identify where they fit and provide more details.
>
> _______________________________________________________________________
> Terry R. Payne, PhD.		| http://www.ecs.soton.ac.uk/~trp/index.html
> University of Southampton	| Voice: +44(0)23 8059 8343 [Fax: 8059 2865]
> Southampton, SO17 1BJ, UK | Email: terry@acm.org / trp@ecs.soton.ac.uk
>

Received on Tuesday, 24 February 2004 03:10:32 UTC