|1 hour60 minute
|November 2 2020 19:00 GMT
|11:00am PST/2:00pm EST
|7:00pm GMT/8:00pm CET
IAOA Semantic Web Applied Ontology (SWAO) SIG
Meetings are normally on the first Monday of the month at 2pm Eastern Time.
- We use GoToMeeting for these meetings - details at Connection Details
- 1 Agenda
- 2 Proceedings
- 2.1 Post for IAOA Website
- 2.2 Ontology Articulation Considerations
- 2.3 Outcomes (redux)
- 3 Attendees
- 4 Next Meetings
- 5 Previous Meetings
- Writing a post for the IAOA website
- Ontology Articulation Guidelines
- New activity focusing on best practice for the semantics aspects of ontology development
- Next steps
- Next Meeting
Post for IAOA Website
This should be a standing agenda item each month
They want something from us each month.
See iaoa.org for examples
They are gently reminding us they would like to see something every month. We did one a few months ago.
There is a set of articles at lower left - it only shows the most recent so our last one has been bumped off.
What do we need to write?
Just a few lines.
Would have a link to web page? Yes
What page" The meeting page? Or the general wiki page for this SIG that all the meeting pages are linked off from.
So that is the link that will go there.
What do we want to write?
What we've done recently? Ontology Articulation Guidelines.
ToddSchneider: Articulate - express (an idea or feeling) fluently and coherently
TS: Do we want to call it that? Articulation has a linguistic meaning and a more common meaning of articulate (see below)
MB and KB feel this is the right notion.
MB: is about articulating the distinctions among ontology, as a pre-requisite to e.g. having design guidelines for the relevant kinds of ontology. hence why we did not want to call it simply Ontology Design Guidelines.
What did we write last time?
The Semantic Web Applied Ontology (SWAO) Special Interest Group will be embarking on a new project to develop a guide for ontology development. Though there are existing guides to ontology development we feel there are gaps in the level of detail of ontological analysis. The intent of this project is to create a guide that will be easy enough to use, but still provide a greater measure of analysis. The goal is to provide more robust ontologies.
So the focus of the earlier description was more about enriching ontology development (design) guidelines, in contrast to what we are saying today.
So we could add something on this, to show some progress maybe?
Other relevant things for the Post:
See latest post from John Sowa on something in itself versus something in a role - important points
TS: This is relevant for this ontology development paradigm.
The IOF had an interesting conversation on that last week, where we learned how BFO approaches those kinds of problem.
ToddSchneider: From John Sowa's email of 31 Oct. 2020:
Basic principle: before classifying any term in an ontology, always ask how it is used.
Advice: Always keep a good dictionary handy when you're doing ontology. Professional lexicographers write them. I have far more confidence in professional lexicographers than in amateur ontologists.
In fact, when it comes to evaluating definitions, I have more confidence in those written by professional lexicographers than I do in most of those by professional philosophers. I have collaborated with members of both professions.
Specialized Terminology Usage
At IOF there was a discussion of the use of words in BFO that did not map to the usual human understanding of these things.
Similarly you would learn words like 'Mixin' to understand OntoUML.
KB to earlier agenda point the meaning of 'multi'.
Ontology Articulation Considerations
IOF use of BFO as the Foundational ontology for IOF. The understanding was that there were some issues with BFO and that if those issues became an impediment we would go to a 'Plan B'
IOF Plan B
Was there a specific plan for Plan B?
RR had recommended Guarino, Guizzardi and others to join that effort, which did not happen. Was this Plan B? Would have been a more neutral and fair Plan B.
BFO has known issues, e.g. being unable to represent an electromagnetic field, energy, quantities and so on.
Would encourage this group to do something to help with these matters.
Is this something this group should add to its agenda?
RR recommends that as part of IAOA we should pursue this, to ensure fairness for all.
This SIG deals with cross-disciplinary matters i.e. Semantic Web versus Applied Ontology. Is this within our remit?
KB, TS agree this would be appropriate. We are attempting to influence different communities.
TS: Relevant to the guidelines (our main agenda today). One thing we might help with is providing distinctions, characterization, assumptions they were created with. e.g. DOLCE cognitive bias, BFO Realist commitment and so on. Both BFO and DOLCE are very explicit about those. UFO it's less clear (MB can look into that). What others?
Should we mention SUMO?
MB we should try to characterize SUMO whether we like it or not - it's not about approving or not, but about characterizing.
TS: The Semantic Web gang are looking to engineer artifacts and need to understand these distinctions.
So we need to be able to describe the characterization so that SemWeb people can consume that information and its implications.
RR: Different communities may create their own foundational ontologies - there are many efforts out there. These may or may not conform to the kinds of characterization we are talking about.
So e.g. the IOF Plan B - given the dominance of BFO in this, neutrality is important so that this (and any TLO) does not become the de facto standard for an entire industry.
So the IAOA should take a role in resisting / preventing such imbalance, and make users aware they can create their own TLO.
RR: People interested in ontology should be able to create their own UO without reference to the philosophy
MB: people " " " should be able to create their own UO AND understand the philosophical commitment; they might or might not make philosophical commitments.
So some things might say they are Realist (in a specific way of using that term) and need to define that in terms of philosophy.
RR Example: category of 'mind' could be rejected or mischaracterized in a Realist ontology (or some kind of Realist ontology). Even within Realism, need to further characterize e.g. of what is it Realist (what do you accept or know to be Real). So there are different types of Realist stance.
So there is a decision tree of things people need to know about for different stances (such as Realism) that they elect to follow.
So Realism is not a single over-riding kind.
Realism - various theories may be realist or anti-realist towards specific matter.
So which Realist stance applies to BFO itself?
RR: The broader consideration is, for a given set of things in the domain, what are the implications of using this or that TLO/UO.
TS: If you are going to conform to some FO, there are some contortions you have to make, to make that thing work.
RR: For most people, should be able to do their own without having to fit within one of the existing ones.
TS: What value does the use of a FO provide?
e.g. one feature of Semantic Web is that they reject that anyway, and create their works for particular uses willy nilly, not looking at specific reuse or extensibility.
TS implies 2 points in the description of SemWeb:
1. The local use case / usage context
2. The time focus - have something immediate for an application but not necessarily maintainable
(which is an issue with software dev generally)
Understanding the role of use cases is another of the things on our roadmap
Conceptual modeling v OWL - there are not really any good tools other than OWL and yet OWL is the wrong tool for that.
Same with UML representations of ontologies.
Our understanding of what it means to be an ontology...
Wrap this up:
1. What needs to be written for the IAOA
2. Also today's notes provide a good jump-off point for sorting out the different dimensions of the problem of describing different kinds of thing that are an ontology. This our main current research topic.
3. We should then be able to take the outcome of these conversations, and feed them (somehow) into things like ISO, IOF, other industry efforts.
(MB: whether you are using an UO versus whether you are using a published UO like BFO or DOLCE)
Further Discussion on Foundational, Upper etc. ontologies
Common Misconceptions - e.g. when and whether you need to use an UO; also how you use it.
Likewise for Foundational and TLO / UO.
Lots to unpack there for another time.
Also what we mean by Foundational, Top level, Upper.
Also the architecture of the ontology framework, whether it has Foundational and so on. Versus e.g. Generic ontology. Implication of a particular architecture by the use of some of these words.
Separately, architecture may refer to the architecture of ontologIES plural, i.e. how you fit together a coherent set of Ontologies across a broad space
There is a range of architectural considerations.
1. The post for today's update
- to be sent to the 4 of us to look over
2. Start to put a wiki page together (MB, under Projects)
Notes in the report in (1) that we are now doing this
and (3) once we have some material in that page we can start to link to it in order to influence conversations elsewhere.
On (3) we can also write a paper once this conversation is further along.
- Robert Rovetto
- Mike Bennett
- Ken Baclawski
- Todd Schneider