[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: RE: [soa-rm-ra] Finalised definitions list and comparison table
Michael: And you have misunderstood my comment – we need more semantic engagement! J The fact that you refer to “children of this definition” highlights the problem: The fact that the text says ‘An actor is a participant’ should not be interpreted to mean that ‘actor is a child of participant’ – that is a very formalistic use of ‘is a’. Maybe the wording needs to change to avoid ambiguity – I’m totally in favour. In this discussion, the model as ‘documented’ in figure 5 is more accurate than the accompanying text, I agree. So let’s please reword the text rather than having a futile discussion about the inconsistencies! Peter From: mpoulin@usa.com [mailto:mpoulin@usa.com] I am afraid, Peter, you have misunderstood what I am doing (this, however, does not mean that I do it right and you get it wrong) So, I do not have a hammer and I do not model 'Participant'. The only thing I do is I take existing definitions (the latest version) and list all terms mentioned in them (with no particular order) as children of this definition. The definitions specified in your spreadsheet in the very left-hand column I interpret as basis terms. They even grouped on my diagram in the same groups as your sheets. Sometimes I may be disagree with the definition but I do not change it, I only note it for myself to tall about it separately. It is a lucky coincidence that I have seen a cycling references for Participant, Actor and Dependent because I do not do any analysis. It is not about wording inconsistencies or definitional errors - I see that Actor references Participant while Participant reference Actor, and I ask the question - is this correct? Probably, there are other loops but it is not my task (at least, now) to find the correct definitions or models. I'm just recording what is said (certainly, XMLSpy and XML Schema is not the tool for modelling our objects). - Michael -----Original Message----- Michael: When you use a hammer, you look for nails – unfortunately, XMLSpy and similar modelling tools cannot handle multiple inheritance and cross-hierarchy models. Participant is a difficult case to model. If you think there are wording inconsistencies that lead you to believe there are definitional errors (and you may be right, particularly in the use of ‘is a’) and consequently circular definitions, then please propose alternate wording for delegate, actor and participant. Peter From: mpoulin@usa.com [mailto:mpoulin@usa.com] @ Peter for the point 1 - I do exactly as you said in my diagrams; there is not problem at all With regard to point 2, I'd prefer to deal with the composite definition as the whole one, without splitting them into separate words; for example: if the definition of 'Peer Social Structure' does not refer to 'Peer' , I would ignore all relationships with the term 'peer' despite its presence in the name. IN other words, if I do a search against this term, I ignore all cases where words, peer, social, and structure appear separately. @Rex I hope that TC addresses the question: should we define all words we use or English is till valid language for expressing our ideas? :-) I can imagine how sticky the Ontology Summit might be but we are writing to not-necessary-ontology-people. If we start define things like 'fact', 'evidence', or 'listener', we risk creating the situation that nobody would talk to us or read the RAF because people loose the confidence in every word they read. In several cases, we re-define terms that have nothing to do with service orientation or architecture. This is what I am afraid of and talking about. I have attached an example of the diagram I am drawing where you can see that one basis term is defined while another term references to the first one. - Michael P.S. In the diagram you can see the basis definitions (as immediate children of the element<<Defined the same in both>>: Actor, Participant and Delegate. While I assume there should be some dependencies between these terms, I have not expected two loops of definitions: Actor-Delegate-Actor (marked by red arrows) and Actor-Participant-Actor (marked by orange arrows). Well, at the end of the day, the purpose of this diagram is exactly this - find discrepancies in the definitions. BTW, you can see elements in the basis definition marked with the version dates and explicit texts of these definitions (sorry, in the given picture these texts are not really readable) If you want to see the diagram I am drawing, you can use any XML Schema visualisation tool, like Altova's XMLSPY and ask me to send you the Schema's text. -----Original Message----- For clarity, value is often in the eye of the beholder, and we need to keep in mind that the one thing that I have heard from the constituencies I work with is that they like having definitions because it keeps them grounded. As fates would have it, I'm co-champion of the Values and Metrics Track of this year's Ontology Summit. Very sticky wicket! Hi Michael: 1 – I was asked to indicate, for each concept that we define (= term) whether other concepts are used in that definition. For example, the definition of ‘capability’ (whichever one you choose) uses the term ‘RWE’. The objective was to indicate simply dependencies between concepts defined, no more. 2 – I couldn’t agree more but they are listed as per the TC’s request. This needs to be addressed by the editors of the sections concerned. 3 – I agree but others don’t – it is an issue the TC needs to address – Chris and I cut back radically on the number of formally defined concepts precisely for the reasons you invoke. We need to distinguish those concepts that have a particular meaning and value for our work: I would suggest they must meet both criteria (particular meaning AND value) to be considered for inclusion as a formal definition. Regards, Peter From: mpoulin@usa.com [mailto:mpoulin@usa.com] Hi Peter, I have 'just' 3 notes: 1) I've found only the columns where you specify the line number that refers to the definition in different versions and its occurrences in different places of the document but not "A new column in those sheets with definitions, that indicates the other concepts referred to in the term definition" - it is not obvious that the referred lines belong to other concepts. I would prefer, if you do not mind, having references directly to other definitions in addition to other points in the text 2) I've found it is really difficult to deal with composite terms like 'Peer Social Structure' because there are too many potential dependencies separately for 'peer' and 'social structure' that may belong to totally different contexts. For example, reference to the line 2717 leads to 'peer' but 'social structure' is not even mentioned 3) in general, I find our extended vocabulary a bit artificial and difficult to operate with: we use relatively common words of plain English in the diagrams and text AND re-define them in our special definitions. Since the words are common, the reader may not suspect that there is special ontology/semantic is meant in our vocabulary. I am afraid, it is overcomplicated. Here is one of many examples:
For given example, I would not define term Listener at all, it is clearly understood w/o our definition. - Michael -----Original Message----- Hi: Chris and I have finished a new version of the definitions table, as requested. The attached version includes: - A new sheet indicating where terms are used in Figures – we have “only†(yes, there are more than 150 of them…) included those terms that have not already been listed having a formal definition – we indicate the first occurrence of the term in a diagram as well as (where it is defined) the line number of the definition in the 17 Jan draft - A new column in those sheets with definitions, that indicates the other concepts referred to in the term definition - The ‘unused’ list is now only terms that are really not used at all, anywhere, in the text but may still appear on a diagram We have not yet included the revised understanding of the concepts of state, shared state, shareable state, joint action, interaction, RWE, execution context – as well as our further understanding of the relationship of those concepts to the SOA ecosystem – I will write up my notes from the offline discussions and circulate those later Thursday. Hope this all helps! Regards, Peter Peter F Brown Independent Consultant Transforming our Relationships with Information Technologies Blog pensivepeter.wordpress.com LinkedIn www.linkedin.com/in/pensivepeter Twitter @pensivepeter P.O. Box 49719, Los Angeles, CA 90049, USA Tel: +1.310.694.2278 --------------------------------------------------------------------- To unsubscribe from this mail list, you must leave the OASIS TC that generates this mail. Follow this link to all your TCs in OASIS at: https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php |
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]