View source for Ontology Summit 2018 Symposium/Chat 20180430

= Chat Transcript for the Ontology Summit 2018 Symposium Day 1 ==
Mark Underwood @knowlengr: Standing by for video - Greetings all

Peter Tuddenham: Hi everyone I am sitting in the summit room in DC

Joanne Luciano: Hi Peter, thanks for letting us know.

Gary Berg-Cross: This is the correct URL for the Symposium.

The URL is shttps://bluejeans.com/472899204
it is also on the Symposium page.

Joanne Luciano: Working now. THANK YOU!

Gary Berg-Cross: Recording started Ram is introducing things...

Mark Underwood @knowlengr: Thanks Gary

Joanne Luciano: Is someone able to provide me some context -- why is NITRD relevant to our symposium topic?

janet singer: They are hosting the face-to-face Summit in DC

Joanne Luciano: Thank you. That's helpful! (It's been a while for me, last time I attended NIST hosted.)

Gary Berg-Cross: Can online people hear the questions?

janet singer: Not well

janet singer: Some better than others

Gary Berg-Cross: Yes, the microphones are hanging overhead but spaced out a bit...

RaviSharma: I saw version 5 last night, thanks,

RaviSharma: Ravi

Joanne Luciano: Sound is reasonable, but a little muffled. Thanks for asking.

RaviSharma: thanks to Ken, he did changes before I sent message to him, based on Cory's inputs for section 5.

RaviSharma: Sound is reasonable but not very clear.

Ram D. Sriram: CAN YOU ALL YEAR KEN NOW

RaviSharma: almost but muffled

Joanne Luciano: Hi Ram, it's a little low.

janet singer: Sound is worse??

RaviSharma: Is it possible to increase volume?

janet singer: That's better Ken

janet singer: Yes

Joanne Luciano: bring the mic closer to the speaker

Joanne Luciano: that's better

Mark Underwood @knowlengr: Agree w/ Ravi, thanks for accommodating

RaviSharma: Ken request you to increase fonts and done

Joanne Luciano: sounds like Mike is saying that we need to put "Context, in context" in context?

Joanne Luciano: The sections should not be in the order of the tracks, that's what the website resource is for

Ram D. Sriram: Ravi: Do you have your hand up

Joanne Luciano: no, and I'm not on audio I don't think

Joanne Luciano: brb, i need to get a headset.

MikeBennett: The issue seems to be that we need to go through the material we have from the tracks, identify the findings and issues, arrange them in a coherent order - but we already have a Summary section that appears as if it has done this but has not.

Joanne Luciano: I agree with Mike. Also, I have content that is not yet there.

Joanne Luciano: sound died

Joanne Luciano: sound is back

RaviSharma: @ken, all - there are the following to be considered: 1. The track speakers spoke on their research or interest topics not always aligned with the track. How important are mention of their non-track inputs?

RaviSharma: contd: 2. Yes we need to abstract content but we also need at least a sentence stating that the speakers contents are not individually acknowledged but assimilated in the Communique'

MikeBennett: In the tracks we identified Issues, Insights and Positions. Is the Communique to be framed around all these or only Issues? Is the Communique to frame our positions on things as in previous years, on only Issues as Ken suggests, or something in between these? Articulating positions that were presented without taking 'a' position. Potential solutions.

RaviSharma: contd: 3. Cross track integration was not intended except in overall contextInContext and will now be hard to attempt as time is short and tracks are diverse, but we should attack low hanging fruits as it makes tracks content purposeful to overall Summit goal.

MikeBennett: Ken's answer: for positions, we don't have a position as a community but we should Cite positions people have taken.

Joanne Luciano: What did we learn from all the presentations?

RaviSharma: contd: 4. Contents of tracks should link to sources where possible, such as presentations and published papers. but there is also value to community agreed general non externally referenced content.

RaviSharma: @Ram - Yes

Joanne Luciano: Re: the different positions, can we say that in some contexts some positions are more useful to take and why? What do we want the reader to get out of this communique?

RaviSharma: 5. Issues are opportunities for current and future efforts including operational maturity of solutions, practice, design, development and research.

MikeBennett: Gary sums up well - we should articulate where there are issues on which people in the community take one, two or many distinct positions, so our audience learns from us what those positions are. There is no 'our' position.

Joanne Luciano: +1 RaviSharma

Gary Berg-Cross: Various speakers and members of the community have taken some positions on the various issues.  These can be noted, but are not the POSITION of the communique.

MikeBennett: What I wanted to avoid is putting out a communique that is only a list of 'issues' whatever those are for.

RaviSharma: @all Issues are Opportunities for improvement, these arise or are discovered during the Summit tracks sessions and in synthesis, and result in Communique'.

Joanne Luciano: Who's speaking? I agree with her. It's related to the material I want to include.

Ram D. Sriram: Donna is talking

Joanne Luciano: Donna is talking about use cases providing context.

Ram D. Sriram: @joanne: Yes Donna is asking about Use Cases

Ram D. Sriram: @Ravi: You have your hand up

RaviSharma: @Ram - Yes

Joanne Luciano: The material I want to include has to do with the evaluation of ontology which needs the context to evaluate it for a given purpose (The MITRE then NIST funded research)

Joanne Luciano: In the non-technical area, they use the term scenarios

MikeBennett: @Joanne thank you. Use case is a term very specific to technology development, not everyday English

Donna Fritzsche: @joanne  that was me, I saw some of the material you wanted to include. I would like to help you or add to the material.

Donna Fritzsche: Russel also followed up in agreement with my statements about use cases.

Donna Fritzsche: similar terms - user scenarios, user stories, business requirements, drivers. consuming application requests/requirements

MikeBennett: @Joanne and @Donna these issues are summarized in the Upper Ontologies track material (Section 3) but are not specific to that topic. See also the introductory stuff in Section 5 which covers similar ground. Please take a look at those.

Joanne Luciano: @Mike, will take a look, thanks.  @Donna, thank you. I look forward to working together.

Donna Fritzsche: thanks Mike! I liked how you worded earlier also.

RaviSharma: @all - now that we have some overall understanding, acknowledgement and non track content addressed, we can concentrate on sections of Communique', starting with what should be in the Intro and Background in Communique' version 5.

Gary Berg-Cross: Put tangential topics like category theory under a general issue category like Representation issues.

Joanne Luciano: What are our main take home points? Would it help start on a new sheet, get our message and structure clear and then divide into breakout groups?

RaviSharma: @Ram - I am the reason for narrative style and agree that track contents should be crystallized and then relate them to issues.

RaviSharma: @Ken - I agree that style should be same and content crystallized.

janet singer: Audio has been OK

RaviSharma: @Mike great inputs.

MikeBennett: Thanks Ravi. I think on our Upper Ontology track we had a lot of insights during the conversations around context v content overall.

David Whitten: What do the speakers mean about "Join the Academy" ?

Gary Berg-Cross: The Washington Academy is a local DC group.

TerryLongstreth: Washington Academy of Sciences, Washacadsci.org

David Whitten: Session has started on Open Knowledge Network on healthcare.

David Whitten: (break out session led by "Guest 1") ???

David Whitten: Heavier approach requires more details. RDF++ may be a way to do this.

David Whitten: Entity oriented knowledge bases, used for search "Apps" or computerized Personal Assistants or consumer applications.

David Whitten: Knowledge graphs are like http://caregraf.info

David Whitten: perhaps using vocabulary from http://schema.org/ ?

David Whitten: generic mention of "world wide web" and "semantic web" - still not clear if this is life sciences or medical or genetic information he is calling "health care"

David Whitten: Creating a Knowledge Graph: Terminological Context (provenance etc) Ontological Context, Alignment Context between participants

David Whitten: "Deep/Heavy" means breaking up into microtheories. Situating in Context Space includes specifying time intervals, time points, time periods, roles of participants, microtheories salient propositions.   An Entity-Relationship (?) drawing of a Medium Deep/heavy approach to Object Design ?Production?/?Protocol?

Joanne Luciano: Looks like Microtheories are what I called "semantic components" in my ontology evaluation work. Will have to look at this more carefully.

David Whitten: Publications of groups of statements have provenance as well as statements having provenance.

David Whitten: mention of "Semantic Science of Ontology" - ?google?

MikeBennett: @Joanne I think so - John Sowa gave a clearer explanation on the conversation part of one of the sessions, where it was clearer that a microtheory is rather like what in FIBO would be a Module or a single ontology file within a modularized ontology. That is, the microtheories are not siloed models as I had previously thought but are modular components within a mutually coherent whole.

David Whitten: mention of situate-ing in space using "located in", "is contained in" and "is a part of"

David Whitten: situate-ing in Time using "exists at" and "measured at" relations

Joanne Luciano: @David thanks for taking notes

David Whitten: situating in information ?? creating entity attributes (information,content, entity) having value of a "literal" - no mention of when attributes are inherited or are "defined as equal" to another entity's attribute.

David Whitten: Describing objects as being identified as being a part of another object. Presumably a metaphorical path to define a specific object.

janet singer: Does someone online have mic on? There's some background noise

David Whitten: This path-described specific object has an attribute which might be a capability ?definition? or a role ?definition?

David Whitten: graph is gone. need to review in slides.

David Whitten: now discussing design for evolution. Contrasting to Semantic Net work in the 1980s.  (perhaps John Sowa's Conceptual Graphs)

David Whitten: mention of OWL, not clear how it fits in Evolutionary design process mentioned on slides.

David Whitten: maybe "Guest 1" is a particular mike in room. Now a different person is named as "Guest 1".

Joanne Luciano: Now 3rd OKN Workshop: Horizontal Breakout  Design for Evolution

David Whitten: Might be nice to have introduction when mike changes hands.

David Whitten: Thank you Joanne.

David Whitten: slide mentions "meta data attribute" of a "trust score". (? definition ?)

David Whitten: data in Triples++ ? part of RDF++ ?  which have a limited situate-ing in context space: physical space, time.

David Whitten: Example: CID5280961 "inhibits" GID2100(ESR2) from Pub Med and Pub Chem. Uses context of "source" of fact and "fromDataSet" context as ChemBL.   ? Example might involve a drug or hormone or protein inhibiting the expression of a particular gene ?

David Whitten: also uses a relation "activates" as in proposition: CID57757 (Estradiol) activates the same "object" ?gene?

David Whitten: oh. CID5280961 is also Genistein. (still unclear what classification exists for this "subject")

Gary Berg-Cross: Ref for SIO Dumontier, Michel, et al. "The Semanticscience Integrated Ontology (SIO) for biomedical research and knowledge discovery." Journal of biomedical semantics 5.1 (2014): 14.

David Whitten: A classifier "Confidence" exists but no data what its values may be.

David Whitten: Phone died- away from sound while plugging in.

David Whitten: Yay! I'm back. Computer audio works.

David Whitten: Guest 1 is a new person. Talking about situate-ing in context space, I think. ?Talking about RDF with Google?

David Whitten: someone who developed RDF standard ?Guha? developed some ideas about RDF extensions ?maybe?

David Whitten: Original speaker is Guest 1 now. Talking about Subject-Predicate-Object as RDF Triple : adds "Starts" "Ends" "Follows" and "FollowedBy" these relations are meta about Triple.

David Whitten: Triple is AbrahamLincoln held USPresident (probably USPresidency-office)

David Whitten: "Starts" and "Ends" relations need to be defined as involving a time-period or SpaceTime objects ?  Speaker promises to do this "in this session"

Gary Berg-Cross: Vinh Nguyen from NLM is speaking.

David Whitten: Very lightweight triples. She acknowledges in the slides that the need for formal semantics. Vinh Nguyen says "semantics is unclear", she wants to create an identifier (presumably to identify a particular triple).

David Whitten: Vinh Nguyen is the original female speaker "Guest 1" above.

David Whitten: "Singleton Property" seems to be a way of classifying some RDF triples. using "marriedTo#1" to classify "subject", rdf:sp to classify "predicate" and "MarriedTo" to classify "object" ... not clear if second MarriedTo is a typo for "marriedTo#2" ?

David Whitten: is marriedTo#1 a shorthand for arg-1 of "rdf:sp" ?

David Whitten: why is "marriedTo" not "marriedTo#2" as a shorthand for arg-2 of "rdf:sp" ?

David Whitten: And what does "rdf:sp" mean anyway ?  s?? p?? == s?? predicate?

David Whitten: In singleton property "marriedTo#1" starts 1965-11-22  is this saying "Bob Dylan" starts at that date? ie. a date of birth?

David Whitten: or is that proposition "starting" on a date?

David Whitten: She is trying to prove an RDF approach is superior?   Are we in syntax wars now?

Gary Berg-Cross: If people online want to stop Vinh to ask a Q, put your hand up...

David Whitten: It is still not sure what Vinh Nguyen means by a "singleto property". ? Is this an alternate syntax for some quinary predicate ?

Joanne Luciano: @David It's not clear to me either.

David Whitten: Her "optimal singleton property" seems to use pointers? with a "ds" of "yago" ; what is a "ds" ?

David Whitten: she changed the undefined term "marriedTo#1" into another undefined term "marriedTo?id=1&ds=yago".  she claims this minimizes cost of representation.

Joanne Luciano: @David "sp" must mean "singleton property"

David Whitten: She claims there is a arrow link from the classification "marriedTo" to the second term "marriedTo?id&ds=yago" .  not clear why.... nor what the semantics of this arrow link is.

David Whitten: The "CKG" publishing model is said to be Triple++. (CKG == ?context knowledge graph? )

David Whitten: she wants to use dimensions of meta data : provenance, time, location, trust score, probability on RDF Triples. ?dimensions of context space?

David Whitten: Thanks Joanne. "sp" probably does mean "singleton property".

David Whitten: It's not clear how a publishing model becomes a "reasoning support". (other than the obvious that any data used by a reasoning process is supported by the data)

David Whitten: She wants to avoid reifying each triple, but she wants to have a global identifier for each triple, which could be used as a reification identifier...

David Whitten: I don't have a charged phone yet. Sorry you will have to read my chats. I'm waiting for the phone charger.

MikeBennett: Are you still note taking or are thee questions? Or both?

David Whitten: is "ds" mean data source ?

David Whitten: sorry Mike. I started asking questions about what she was saying. Perhaps I need to start my line with something indicating I am personally commenting. Do you have a suggestion?

MikeBennett: Thx.

David Whitten: my?? maybe starting the line with "my??" if it is a comment rather than a note taking.

pfps: Is there a pointer to these slides?

MikeBennett: or just Q: at the start? I was struggling to extract the questions from the (very useful) summary text. Would also be useful for when someone tidies up and publishes the chat log as proceedings.

David Whitten: my?? I had a hard time understanding what she was saying because I didn't see any definitions.

Joanne Luciano: I agree with David.

Joanne Luciano: Vinh just said that they can't do heavy reasoning with sp

David Whitten: :: Vinh was asked if she was willing to go to the Common Logic syntax. Vinh says there is a lot of information in RDF triples.

David Whitten: :: she says when she has temporal information she can start temporal reasoning.

David Whitten: :: She says systematic categorization makes it easier to share information.

pfps: Is there a pointer to these slides?

David Whitten: :: There is a community, and data publishers.

pfps: This looks a lot like what Wikidata has.

David Whitten: ? Vinh could you explain why you are saying you are replacing RDF reification when you are just providing a reification identifier.

David Whitten: :: someone ?Janet SInger? is suggesting she is advocating a syntax more than a semantic approach tying to the Dublin Core or schema.org ...

David Whitten: :: sorry ?Janet? was suggesting to Vin, not herself.

Gary Berg-Cross: We will ask Vinh for the slides to be posted by Ken.

David Whitten: my?? what is the difference between a group on RDF Triples about Triples  and a group on RDF meta-data about triples. ?? doesn't "meta-data" already mean "data about data" ??

David Whitten: :: ?Janet? also suggested that Vin's approach doesn't look backward compatible to the RDF graph tools vendors.

Donna Fritzsche: @david - that was me Donna not janet

janet singer: @David  I'm just online. Maybe it was Donna if that was someone in DC

David Whitten: :: Mike suggests an attachment to an upper level ontology.  She says she wants to make this available to people who want a light-weight ontology.

David Whitten: :: Mike reminds Vinh that an upper ontology rarely affects application designs.

David Whitten: :: someone points out that if you don't note when an RDF triple has an impact on the truth-value associated with it, then you are affecting reasoning used with that triple.

David Whitten: my?? isn't "RDF reification" something typically supplied by the RDF graph engine?

David Whitten: :: Vinh says she has decided to provide a very limited/lightweight method of representing context.  She says it is general representation method, but they have only represented a small amount.

Mark Underwood @knowlengr: I hate to beat my same old drum, but the domain model will always win out in scenarios like cybersecurity, safety, risk management

MikeBennett: Comment on :: 12:18 I was not saying that upper ontology rarely /affects/ application design, but rather correcting the notion that because I was asking if Vinh was considering referring to any upper ontology partitions, that this somehow implied that I thought the OWL triples for UO partitions needed to be in some or any triple-based application. Upper ontology partitions should /affect/ an application design but you can affect a design without being part of the /content/ of that application. Separation of concerns means you can refer to conceptual matter (including UO) without it having to sit in your solution database or triplestore and slow down some reasoner.


[12:27] RaviSharma: Is there a slide deck to download?


[12:28] KenBaclawski: I will ask Vinh for her slides to post on the symposium page.


[12:28] Gary Berg-Cross: Contextual info is "about" data and so is meta data. So we can call some context provenance metadata. Others source metadata.


[12:28] RaviSharma: the original deck in Symposium is a different smaller version for Vinh and Gary?


[12:28] David Whitten: my:: Vinh said she would make the slide deck available later. (but it seems to depend strongly on some papers as well. Hopefully they will be made available too.


[12:29] pfps: It's very helpful to have the slide deck available now so that we can look back at the slides.


[12:30] RaviSharma: @David - thanks. Ken also some interoperability slides are shown only on recording such as Feb 21 and March 21.


[12:30] David Whitten: my:: Fitbit data is basically personal lab results.  The LOINC standard provides a way to describe those results, effectively meta data about the results.


[12:30] RaviSharma: @ken: Ken also some interoperability slides are shown only on recording such as Feb 21 and March 21.


[12:31] Joanne Luciano: Are we breaking for lunch?


[12:31] Mark Underwood @knowlengr: Ken, thx for hunting down the decks


[12:31] David Whitten: my::  It is 12:31 EST, so I'm not surprised they are breaking for lunch.


[12:31] RaviSharma: @Ken is this a lunch break starting now?


[12:31] TerryLongstreth: Yes, We are breaking for lunch


[12:32] David Whitten: my?? Mark is @knowlengr your twitter id?


[12:33] Vinh: https://www.slideshare.net/ntkimvinh7/ckg-portal-a-knowledge-publishing-proposal-for-open-knowledge-network


[12:33] David Whitten: my:: thank you Vinh. could you share links to your papers too?

[12:34] David Whitten: my:: I'll take a liverwurst on rye. Do you deliver to North Carolina?


[12:34] Vinh: https://www.slideshare.net/ntkimvinh7/www2014-singleton-propertyfinal

[12:35] Vinh: my paper https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4350149/


[12:35] David Whitten: my?? Does anyone who is remote like me want to discuss any of these issues while they are at lunch?

[12:36] David Whitten: my:: by the way the LOINC standard is established by the Regenstreif Institute in Indianapolis.

[12:40] David Whitten: my?? could someone who uses RDF tell me if RDF reification already gives you access to an identifier for an RDF triple?  This seems to be a significant issue that Vinh was discussing.


[12:40] Joanne Luciano: @David, good idea, but need to take care of something and may not be back to my computer until a little after 2. I plan to connect by phone if I can.


[12:41] Mark Underwood @knowlengr: David: Yup


[12:45] David Whitten: @mark. Ok I've followed you @davewhitten

[13:18] David Whitten: :: There is an issue of describing a context in which the ontology is seen This is an internal context or neighborhood of context where the ontology is seen. There is also a context for the data used by an application, perhaps this is an external context.

[13:18] David Whitten: :: This is a lunch discussion, but I don't recognize the voices of folks suggesting these distinctions.

[13:20] David Whitten: :: Perhaps if the boundaries are fluid or changing, this is an area in context space. This may be an idea to develop in the communique pre statements.

[13:21] David Whitten: A Context includes a model. A microtheory says there are salient assumptions that are never contradicted.

[13:22] David Whitten: :: Consistency is guaranteed for all microtheories connected using a genlMt relationship between microtheories in a genlMt hierarchy


[13:22] Mark Underwood @knowlengr: David, currently speaking is Gary

[13:22] Mark Underwood @knowlengr: Then Terry L


[13:23] David Whitten: :: Communique document changes: 7.3 and 7.4 should be tied together somehow.


[13:23] Mark Underwood @knowlengr: This is Donna

[13:23] Mark Underwood @knowlengr: Just before her was Todd

[13:24] Mark Underwood @knowlengr: Steve, we've missed you!

[13:25] Mark Underwood @knowlengr: The erstwhile Brit is I think Mike Bennett


[13:25] David Whitten: :: There is a possible ontology to model contexts...


[13:38] RaviSharma: what time after lunch is estimated to start?


[13:39] David Whitten: :: I believe the restart is at 2:00 pm EST


[13:39] RaviSharma: Thanks it gives us about half an hour. Regards


[13:54] MikeBennett: I'm not all that erstwhile...


[13:55] Mark Underwood @knowlengr: Mike

[13:56] Mark Underwood @knowlengr: The remote audience is getting a reminder of why the IC is interested in ontologies to help with context

[13:56] Mark Underwood @knowlengr: --think listening "devices")_

[13:57] Mark Underwood @knowlengr: David: that's Ken speaking


[14:02] janet singer: Yes clear


[14:03] David Whitten: :: Ram is talking re his experience in manufacturing world for last four decades.


[14:23] Mark Underwood @knowlengr: this is great.. Does the automated notetaker write "Ram interrupting himself with annotation ..."

[14:24] Mark Underwood @knowlengr: Several concurrent discourse contexts in play


[14:25] David Whitten: :: Someone is overtalking the speaker. I don't know who.

[14:25] David Whitten: :: the double conversation is overwhelming


[14:28] Mark Underwood @knowlengr: that is Ram speaking over a video of himself speaking


[14:37] Joanne Luciano: CPM stands for?

[14:40] Joanne Luciano: PSRL product semantic representation language


[14:41] MikeBennett: CPM Core Product Model I think.


[14:47] Joanne Luciano: PSL Process Specification Language


[14:48] Mark Underwood @knowlengr: PSL is important because one of only 3-4 ISO standards that incorporate ontologies

[14:50] Mark Underwood @knowlengr: see also ISO 16739


[14:50] Joanne Luciano: When I did my ontology evaluation framework I had ISO as a use case.

[14:51] Joanne Luciano: I missed "CT" What does CT stand for?


[14:52] janet singer: Category theory


[14:53] Joanne Luciano: @Janet thx


[14:54] Mark Underwood @knowlengr: Joanne, ISO generally, or one of these standards

[14:54] Mark Underwood @knowlengr: ?


[14:55] MikeBennett: @Mark hh:48 comment: PSL also important as one of only a few ontologies that incorporate process! e.g. in one ontology project I've worked on, the process semantics (similar to PSL) were removed as not being recognized by the OWL experts as being ontology content.


[14:56] Mark Underwood @knowlengr: IEEE 1872 standard is "Std ontologies for Robotics and Automation"

[14:57] Mark Underwood @knowlengr: Mike: That puzzles. Does that mean workflow / orchestration cannot be in an ontology for those folks?

[14:57] Mark Underwood @knowlengr: Maybe I need to look at the latest DOL spec...

[14:58] Mark Underwood @knowlengr: Mike.. mildly off-topic... Am trying to start up interest in DHS / FS-ISAC in using FIBO in cybersecurity for finance.  Early stage.


[15:00] MikeBennett: Seemingly so. Or they thought it was some abstract and therefore to them unnecessary 'upper ontology' stuff. OWL is more usually appropriate for stand alone applications so there was an appetite to remove 'abstract' stuff that has no clear role in the reasoner. The process primitive foundational models we did fell foul of that perception. Other ontologists would have reacted differently I'm sure. Looking at PSL later, it is similar in nature and broader / better thought in scope than what we had, but same kind of thing.

[15:00] MikeBennett: @Mark Yay. See previous comment re FIBO.


[15:02] BobbinTeegarden: @Mike nice to see the FIBO effort validated


[15:02] Mark Underwood @knowlengr: Marrying pieces of Kubernetes to ontologies (e.g., DMTF CIM) requires workflow characterization in depth

[15:02] Mark Underwood @knowlengr: Also needed for attribute-based security (!)


[15:03] RaviSharma: Is Ram the speaker now?


[15:03] Mark Underwood @knowlengr: Ram is moderating, Mike B was discussing FIBO


[15:03] BobbinTeegarden: Does all this imply that you could at some point 'execute' your ontology?


[15:04] Mark Underwood @knowlengr: Ask Cory 


[15:04] Donna Fritzsche: https://en.wikipedia.org/wiki/Process_Specification_Language


[15:05] MikeBennett: Flora2


[15:05] Mark Underwood @knowlengr: https://en.wikipedia.org/wiki/Flora-2


[15:05] BobbinTeegarden: Isn't FLORA more rules level than process?


[15:05] Donna Fritzsche: http://flora.sourceforge.net/


[15:05] Mark Underwood @knowlengr: Could somebody recap that rationale for FLORA here?


[15:05] Joanne Luciano: @Bobbin The spectrum that Ram had next self describing then self executing (if I recall correctly)


[15:06] Mark Underwood @knowlengr: @Bobbin - My personal take is that translators / build tools are a better models, but microservices could certainly execute pieces


[15:06] BobbinTeegarden: @Joanne Thanks, right.  Did we get there yet? ;0)


[15:07] MikeBennett: Isn't that how the Terminator movies start?


[15:07] Mark Underwood @knowlengr: Mike: they do in *my* mind


[15:08] RaviSharma: @Pete - Nice to have you (Peter Patel-Schneider) as a participant, welcome  regards


[15:08] Mark Underwood @knowlengr: All- sorry to miss the F2F collaboration this year!


[15:09] RaviSharma: Ravi meant to type @Peter


[15:12] Mark Underwood @knowlengr: less discussion of Hadoop because it is commoditized

[15:12] Mark Underwood @knowlengr: Moved to Docker, Kubernetes, Spark, Mesos, etc


[15:12] RaviSharma: @Ram, HDFS is Hadoop used as assumed platform for all Big Data is it not?


[15:13] Mark Underwood @knowlengr: Russ is a good resource for that material


[15:13] RaviSharma: Hadoop file system for ingesting and dealing with big data processing.


[15:13] Mark Underwood @knowlengr: Usually "data lake" means Hadoop.  (The managers have taken over the project.)


[15:14] BobbinTeegarden: Where are the slides from the previous presentation?


[15:14] RaviSharma: I thought data-lake was a concept like datamart for big-data?

[15:15] RaviSharma: @Todd what happened to your logging in chat

[15:16] RaviSharma: @Todd it is gone now.


[15:16] Joanne Luciano: @Bobbin where are you (geographically speaking)?


[15:19] BobbinTeegarden: @Joanne  back in grey Bellevue, there in spirit!

[15:26] BobbinTeegarden: Has anyone actually implemented PSL?

[15:26] BobbinTeegarden: @Ram would you share your slides/video?


[15:27] RaviSharma: @ram where is the link?

[15:27] RaviSharma: @Ram

[15:32] Joanne Luciano: @ToddSchneider10 things could be worse 


[15:37] Gary Berg-Cross: Can online people hear the discussion or do we need to pass the mike?


[15:37] janet singer: Hear chatter but don't know what you're talking about so can;t follow


[15:39] RaviSharma: Domain related: 1. Role of Context and examples were presented for only a few domains such as healthcare, sub-domains within healthcare, and financial services.  2. Context is often implied and understood sometimes explicitly at Meta level or as metadata about ontology/application 3. Context congruence or synchronization at subdomain level implies compatible values for each sub-domain related contexts. intra-domain contexts have also to mean the same (Synchronized-in spatio, temporal or parametrically) compatible values for contexts to meaningfully apply inter-domain ontologies or applications. 4. Domains have fuzzy or often overlapping areas expressed for example in terms and Vocabularies. more to come on domain track.


[15:40] Duane Hybertson: 7.5 refers to two types of context: compositional (part of a larger thing) and metalevel or category of a thing. This is an important distinction. Are you deleting this section?


[15:41] MikeBennett: There is a broader idea that was mooted over lunch, which is that the overall structure of 7 can be divided into two broad headings (contextual stuff within ontologies versus ontologies as seen, used or created within some context). Then some of the 7.xx sub headings are level 3 headings under one or other of those 2 (or 3. relations among these 2).

[15:42] MikeBennett: So 7.6 seems to be a possible sub heading of the right hand side of this dichotomy.


[15:45] Ram D. Sriram: @ravi: Sent a link to Ken. He will put it on the web


[15:45] RaviSharma: @CQ -for Communique shortform  ref Page 2 first para continued  to also include :the flavor of
context include circumstances, conditions, factors, perspective, scope, state of affairs,
situation, background, scene, setting, and frame(s) of reference, Parameters and metadata about the content.


[15:47] Gary Berg-Cross: Ontologies and services they support are limited are limited to the scope of knowledge. Their "competence" in and knowledge about things is limited to a particular context within which the ontology was developed.  This scope may be detailed by competency questions used to design the ontology.


[15:52] RaviSharma: @ram thanks, also heard Sub and Ira presentation recording, thanks for supporting great grammar related extensions to NLP through NIST.


[15:52] MikeBennett: Based my summary just now in words (also written above at hh:41), I would suggest we can look at each of the current 7.n headings and ask ourselves is it a sub heading of 7.A, 7.B or 7.C, if those were the new numbers for the 2 concerns mentioned and the interaction between those being 3 (I've called them 7.A / B / C here but they would become 7.1 / 2 / 3).


[15:57] BobbinTeegarden: @joanne  'shared (prior) knowledge' = implicit context?


[16:02] MikeBennett: Suggested that perhaps Perspective can be defined usefully as the kind of Context where the W is some Agent (a Who); that is the perspective from which something is interpreted  including some ontology but actually some anything. Also per Gary, Agent brings their own knowledge to their interpretation of a (ontology / model / assertion or whatever). This is their background.

[16:03] MikeBennett: So per Cory this is one kind of matter under 'the Context of Ontology' as distinct from the Ontology of Context??


[16:03] RaviSharma: @CQ - Section 7.11 Note: While time is likely to be part of a context, it is also often in the ontology. It
is not clear that time is an important context issue by itself. In any case, there is
no reference to this issue. - Time is often embedded in ontologies but for meaningful interpretation of content, ontology etc., one needs to align or synchronize time as metalevel parameter in the context for them to logically transfer information inter-domain or for general interoperation.


[16:04] BobbinTeegarden: @Mike that 'background' is an implicit context -- another kind of context... (as opposed to explicit context, the place of current focus)?


[16:04] Gary Berg-Cross: a perspective may involve background knowledge.


[16:04] MikeBennett: @Ravi this is a good reason why many of the things that might sometimes be the context /of/ some ontology or other model, can also be made explicit as context /in/ the ontology or other model. Time is a good example. All the Ws are.

[16:06] MikeBennett: So all things we can think of as being Context (all the Ws), can either be /in/ or /of/ a ontology. Nothing unique about Time in that regard.

[16:07] MikeBennett: Who What When Where Why hoW.


[16:08] David Whitten: Didn't Pat Hayes write a Naïve Time Ontology?


[16:09] MikeBennett: Time is comparable with other Contexts types, but it makes a good thinking exercise because we experience it so differently.

[16:10] MikeBennett: Also on our track we had other kinds of context not being Ws, namely granularity/scale, process and system (Smith, Track B Session 1)

[16:11] MikeBennett: Situations and settings combine time and place to form a compound kind of context (Gary comment).


[16:14] RaviSharma: I like how mike described these and we may want to say about alignment or synchronization of context to be meaningful and logical e.g. spatio-temporally or for spectrum if we are talking of frequency in science and engg.

[16:20] RaviSharma: CQ: 7.16, context interpretation is also influenced by picture theory of meaning, namely model we make in our mind for meanings, so Escher is very confusing for a logical geography oriented mind yet it is a wonder in terms or art but far removed from reality.

[16:22] RaviSharma: @all are we discussing context Provenance? and context issues based on PROV?.


[16:24] MikeBennett: @Ravi exactly - broader metaphor for how any artifact is seen within some context - applies to picture, data model, other model, ontology etc. Might make good introductory material on these distinctions if we use that.


[16:25] RaviSharma: @mike, I like your description for 7.16, it carries sense, we ought to keep it and not drop it


[16:25] Duane Hybertson: Why not include the figure-ground construct in the Perspective section?


[16:25] RaviSharma: @Mike - i meant no typo.


[16:27] BobbinTeegarden: M C Escher is a great example of using contextual 'jokes' by finding interest/humor in mixing contexts (scoping in and out, impedance mixmatch in perspectives...)

[16:29] RaviSharma: @CQ - 7.18 it is relevant as NLP leads to discovery of context as well keeps track of dynamic context changes (e.g. one text story itself)


[16:29] MikeBennett: 7.1 we decided is not really an introduction to context but is a sub heading in its own right, containing stuff.


[16:30] Donna Fritzsche: agree with @mike Bennett

[16:31] Donna Fritzsche: @bobbin - good point about contextual jokes!


[16:31] TerryLongstreth: @Duane,@bobbinTeegarden - Thanks for your thoughts. I'm mulling over the idea of expanding them in a paper. longstreth@acm.org


[16:37] MikeBennett: We need to decide whether the flow is general to specific or specific to general. I was assuming the former but it's right that we ask that question.


[16:37] Robert Rovetto: Going over the Communique document (latest version), might I recommend two things. (1) Include more foundational ontologies in section 3, p.4., such as Cyc, SUMO, UFO, GFO, PROTON, KRR Lattice (Sowa), etc. The Communique should reflect a broader upper ontology world. (2) Use the term 'Generic Ontology' Or 'Foundational Ontology' rather than top-level ontology, because the latter assumes a certain architecture/approach (levels), but not all developers and ontologies use that architecture/approach.


[16:37] MikeBennett: Then eg upper ontologies are one (specific) kind of solution to the broader problem of representing Stuff. Where Contexts are a kind of Stuff.

[16:38] MikeBennett: @Rob agreed, these are listed in our slide deck but not in the material I wrote for the blog page. We should have these somewhere.


[16:39] RaviSharma: @CQ - 6. System and "systems of systems" is wider use and engineering is a specific implied spin therefore domain and interoperation merging makes sense but it does not easily combine with 6.


[16:42] MikeBennett: @Rob hh:37 when we used the terms Upper ontology or TLO in our track on that topic, we meant only that exact thing as defined in the literature, and not the broader range of interpretations of the words, that you are referring to, so I believe it's appropriate to retain the wording we have.


[16:48] janet singer: My microphone doesn't seem to be working, but Jack and I had asked that the SE section be deleted as  a stand-alone track.

[16:50] janet singer: Appreciate that Todd added MBSE paragraph, but maybe that can be folded in elsewhere


[16:59] MikeBennett: defining something is not the same as approving of it. Indeed, one needs to define a thing clearly if one is to say that one does not approve of it!

[17:01] MikeBennett: So Track B took a clear definition of TLO in order to discuss the merits and demerits of those in modeling context.


[17:03] janet singer: Yes, Donna - that's it. Its an early email from Jack that he didn't intend to be in the communique.

[17:05] janet singer: Jacks point re context and environment was also deleted (7. Something)


[17:09] Mark Underwood @knowlengr: "Environment" has a number of specific meanings in the ACM Computing Taxonomy. ... There is value in doing some kind of crosswalk, even if a footnote


[17:10] janet singer: I think a good solution is to do what Ravi did not want to do: put Todd's paragraph on MBSE in the section on Domain Specific needs (together with anything else you want to add)


[17:10] Mark Underwood @knowlengr: https://www.acm.org/publications/class-2012


[17:12] janet singer: SE as context of ontology development, use, etc. is too multifaceted for a paragraph


[17:13] Mark Underwood @knowlengr: Soapbox sidebar: The MBSE references among the few that can be picked up by the IDE builders. There is work here that will enhance the ontology community;  MBSE reaches a broader community where models have different, but not ontological shape


[17:13] janet singer: Yes  SE, ontology and context could be topic for next years summit

[17:15] janet singer: A lot of these SE/Ontology/stakeholders/context issues topics were briefly raised in our track.

[17:17] janet singer: @Todd  Jack would see construction, architecting as cases of Systems Intervention (with SE and Engineering of Systems as other cases)


[17:17] Mark Underwood @knowlengr: Janet - I'm pondering it for a book topic if you would like to discuss offline


[17:18] janet singer: Remember that our topic was Harmonizing Diverse Conceptualizations in Multi-Context Systems Engineering

[17:18] janet singer: @Mark - yes.


[17:18] ToddSchneider: Janet, I don't have great knowledge is those areas, but the problems have similar practices for practical reasons).


[17:19] janet singer: We could barely scratch the surface

[17:20] janet singer: That's why further work is warranted  to cover all the important issues that have been raised


[17:24] RaviSharma: @all, NITRD, Ram, Ken thanks for a good first day. Regards, Ravi bye for now.