Take aways from the IOF Face to Face meeting in Feb 2024
Initially populated by Evan Wallace, but please add your own thoughts. Melinda’s have been added.
General
Significant emerging manufacturing areas are using or interested in IOF ontologies and knowledge graphs. Communities represented in Tempe included: BioManufacturing, BioPharma, and Additive Manufacturing. (Obs, +Com, +Cont)
NIIMBL is becoming a part of IOF and will be contributing ontology content
Q - Would the Additive folks do a new version of the CDT in IOF?
New US Government communities using BFO-based ontologies
US Customs and Border Patrol (DHS/CBP) has major efforts using ontologies and knowledge graphs for data integration for decision support. US DOD has signed a statement regarding department wide use of BFO and CCO for ontologies that they develop or use. (+OntUse)
ISA-88 (batch process control) and ISA-95 (enterprise – control system integration) standards are being used by NIIMBL and investigated by the IOF PPS WG. (Trnd, Cncn)
These standards were developed predominantly by and for process and continuous industries and have some quirks in their models as well. Subsequent work has added elements to ISA-95 to address many constructs/notions needed in discrete domains.
What about notions needed for control in discrete industries? Maybe MTConnect can fill this role in IOF?
CCO now has a Governance Board. Rod Rudnicki has retired. CCO is under review to become an IEC Standard.
From Architecture Task Group sessions
There is still a gap to bridge from ontologies and knowledge graphs as a research and bleeding edge technology to ontologies and KGs as a robust and mature technology for use in enterprises. Some examples discussed in Tempe, AZ include:
Available and affordable (including free) DL reasoner software is not presently being maintained. IOF action: We should work with other OWL user communities (e.g. FIBO, IDMP, ISO IDO, and US DHS/CBP) to advocate for better support (see https://oagi.atlassian.net/wiki/spaces/IOF/pages/4882432002). (Cncn, Act)
New practices are needed for enterprise graph database and application developer/implementations to mitigate change from evolution of IOF ontologies when used as a “schema”. (Cncn, Act)
IOF should strive for “right sized” IOF reference ontology modules. (Pol)
That can be published in a reasonable time
That are more easily reused by multiple domains or subdomains where applicable
The IOF development process documentation needs to better communicate the expected scope/nature of competency questions and test data that are required/developed as part of the development process, and how this relates to the scope of an individual ontology where such competency questions are used. For instance, it is expected that ontologies that extend the ontology being developed may be needed to answer competency questions against sample data. (Cncn, Act)
From Barry Smith’s session - change to BFO
Time
The BFO implementation in OWL names “bfo-2020-without-some-all-times.owl” will now be names “bfo-core.owl” and will consist of all BFO classes and relations identified in ISO 21838-2. No object property IRIs will be changed.
The rdfs:label for all “some time” relations in the OWL version of BFO will be updated to remove “some time”. The “all time” sub-properties of “some time” relations will be removed from bfo-core.owl.
Temporally ambiguous binary relations will be implicitly interpreted as holding at some time, because it is a formally weak interpretation. A skos:scopeNote will be added to the resulting relations directing those interested in more robust semantics for time to guidance and the use of Temporalised Relations.
Previously conformance to BFO ISO required use of the Temporalised Relations. Acknowledging the different needs to BFO stakeholders, conformance to BFO ISO will no longer require such a robust representation of time. Though users who need such representations are encouraged to use Temporalised Relations as a temporal extension of BFO.
The Modal Relation ontology is a copy of the relations in the Core BFO Ontology but used for things (such as products or processes) that do not exist yet but are described in documents such as design specifications. (Q - is this something IOF is going to look at?)
Characteristics
Processes have characteristics, not qualities. Qualities are restricted to continuants. Qualities are SDC and can be realised (but might not be). Examples of characteristics include velocity, gpm etc. One can only have a characteristic if a process exists. (Q-modelling industrial processes needs this concept of characteristics - where is this in IOF/BFO now - and are there plans to include it?)
Software and patents and roles
Software cannot have functions in BFO. Only material entities can have functions (and dispositions). However software obviously has a function and realizes that function when it is installed on a computer. (Q-this seems like a contradiction to me)
People do not normally buy software, the buy the right to use it. What are rights. Barry is working on an ontology of legal entities. Laws, rights etc are all GDCs.
What we manufacture when we manufacture software is a copyable pattern = a GDC.
Examples of these include mathematical models, blueprints, knitting patterns, musical scores, laws, rights, permissions, contracts.
GDC’s cannot have functions (because they are not physical) but they can have roles.
Pfizer designs a new molecule which they have not synthesized. They then patent the molecular pattern. A patent is an GDC which confers on the patern a new role. The role is (roughly) inherited by the molecule that is eventually synthesized.
When you buy the rights to use a given piece of software, you and the software (pattern) acquire a new relational role. This is analogous to the standard relational role of legal ownership. But standard ownership is a cluster of rights (to use, to sell etc). So this is not correct for software as you are buying only one right, the right to use. This issue of rights is an active area of work for Barry. He made a 30 min video of this back in 2016 https://www.youtube.com/watch?v=WIJJlnLLWGU
Barry suggests that IOF will need to work with modules pertaining to: ontology of law (ownership, rights, insurance ..), business processes, and commercial exchange.(Act)
Qualities ontology
A qualities ontology extending bfo:Quality was created by the Materials WG for some of their initial work. We discussed using this as a starting point for work in a domain independent working group (either Core or a new WG) to develop, essentially, a library of bfo:’quality’ and bfo:’process characteristic’ subclasses needed by all domains (which would be the IOF analog to the International System of Quantities formally described in the ISO 80000 series of standards). We also discussed the desire to keep this separate from the Core ontology to keep Core small. IOF needs to determine how and where to do this work. (Act, Pol)
The slides provided a link to this work but I can’t get the link to work. It is https://github.com/iofoundry/ontology/tree/'qualities'
The work is based on using classes in the PATO ontology. Currently the qualities ontology has 991 terms, of these 857 are from PATO.
A small group elected to work on this with the Materials WG including Todd and Melinda. Melinda’s interest relates to work on a Failure ontology - looking at modelling data describing failures of processes and things where the data comes from design (potential failures) and operations and maintenance (actual failures).
Tags for categorizing bullets above
Tag | Description |
---|---|
Act | IOF Action needed |
Cncl | Conclusion |
Cncn | Concern |
+Com | New IOF member community |
+Cont | New potential IOF content |
Obs | Observation |
+OntUse | Increased ontology use |
-OntUse | Decreased ontology use |
Pnt | Point |
Pol | Proposed policy |
Q | Question |
Trnd | Trend |