Proceedings 20150803 FIBO-FCB FCT.docx
20150803 FIBO-FCB FCT
Today we will look at the draft specification.
Currently this has placeholder material from IND for most of Section 10 but this will change.
Section 9: Participants will then be able to review the content by way of the diagrams and definitions. Meanwhile we are seeking reactions on the new presentation arrangement.
Ian: this looks great.
Agenda: See Slide 2.
Updated Word doc and PDF will be posted to the wiki. 13M and counting. Meanwhile, comments via separate notes please, on sight of the PDF version.
Status of the ontologies also updated today. 2 more ontologies to be uploaded to GitHub. We need more review. The Word version will be at the bottom of the wiki page when it is uploaded, and Elisa will mail everyone when it's there. In the draft now: stuff we are proposing to add to Foundations, and Financial Instruments. Not there yet: the other material. Arrangement (classifications etc.), Quantities (new module) Accounting etc. for Foundations, are available in GitHub for addition to FND Pink. Currency stuff is now added to the existing Currency Amount ontology in FND. Codes are in GitHub including currency codes, and are in the LCC namespace. Products and service, payment and schedules, are in Foundations. To be updated today or tomorrow in GitHub. Everything else is up to date in Github in Foundations in pink.
Language sand Countries see slide 6. Now in LCC namespace. Also in Github in Pink
Changes to reduce the reasoning overhead has worked well (see previous JIRA on
Slide 7 outstanding actions. See meeting notes on the wiki for the Summarized on this slide. Slide 7 is up to date for open actions, except MB has delivered the thing in bullet 4.
JIRA Review: Slides in the wiki include the diagrams for FBC content that will be loaded into the document (June 17 parts 1 and 2 and the 29 June slides). Some refactoring e.g in Transactions, and the stuff we moved to Foundation. Where the slides differ from the draft spec, the spec is ore up to date. As re the ontologies in GitHub. The ontology that is at risk for Aug 24 deadline is Business Centers. That will only be needed for OTC Derivatives (FpML term). Elisa now filtering out those products and services concepts that are now in Foundations. Will put these in a new Module. Example individuals not in Pink. Includes test individuals for Citi. Some things to be updated to reflect the move to LCC in the next 48 hours. See Draft FBS Specification Materials page in the wiki for the latest spec.
FBC16: will be done this week
FBC8: received the email this morning
FBC4: reconcile BE soon and ongoing
FBC2: not agreed.
FBC-12: Gareth looking at
FBC20: Gareth looking at. corresponds to actions in Slide 7.
High level services has a list from Steve Creek, should be sufficient. Nothing for investment banking at this point. Or for fund management companies. So may not be definitive yet. However the goal was to test relationships with some individuals, which we can do with what we have. So Business Centers is the main outstanding thing
Homework: mostly covered in the above. (slide 8). By tomorrow night (Pacific) people can download the latest versions of everything from GitHub.
PartyRoles pattern for buyers and sellers what we currently have is in draft spec. to review next week. There is also a section in the spec for use cases would like more on this, including examples. Can be added later if necessary but ideally we should have it by Aug 24. Elisa will reach out to the banking members of this group for more on that.
Review of changes to move things to Foundations.
Slide 9 show what items have been moved to Foundations. David Saul: I will get some of our real banking SMEs to look at it. Some things left off which depended on business entities, also regulated commodity which has dependencies on FBC definitions for regulators. So those stay in FBC. In the original FBC ontologies, some of the classes are embedded in the Functional entities ontology and need to be moved to the new, small Financial Products and Services ontology to parallel what is now in Foundations. Then have the financial services entities depend on the new versions of financial products and services. Not present in the spec: Payments and Schedules. That will be done today. Don’t' expect a new version of the doc just for that section. Not sure about moving financial services entities. See wiki again:
Diagrams and Screenshots from Protege see Draft FBC Specification Materials wiki page. To be completed by end of this week. Please review material this week.
Slide 10 document status. Please focus on Section 6 onward. Section8 is not all boilerplate will include the namespace abbreviations unique to this spec as well as dependencies. Next effort will be Section 9. Where are the individuals and examples going? The individuals that are jurisdiction specific, there are classes in those ontologies as well. This will be part of the spec, probably in an annex. What does Pete recommend? Non jurisdiction specific stuff first e.g. legal entities types (individuals (representing things like DTCC, Citi etc.). These are non normative and are examples only. We may but these informatively in a non normative annex. The ontology itself, we need to determine if these are normative or not. Pete recommends we include example as part of the normative spec itself.
Given we have done the work on examples, and Pete has founf them valuable for
reviewing and understanding the ontology itself so the will be useful. Need to determine how to include these in the normative parts of the spec. Does this include named organization? Even though these are in the public domain, it would make sense to speak about how the examples would work without saying real things about those entities. Pete recommends using real entities so people can compare with published sources. If we do that we need permission from those firms. Given there is nothing but publicly known facts about e.g. Citi, can we have permission from Citi?
Action: Determine how best to incorporate the examples in the body of the spec from public information about the banks.
Action: Mike B to review Section 0 edited. 0.3 submission team updated. 0.4 minor language tweaks. The basis of this text is IND. MB also to review 0.4.2
0.5 updated with examples. Section 1 Revised with specifics. 1.2 specific to this spec. Please can people review Section 1.2. Section 1.3 additional sources added for definitions.
Action: Pete to review Section 2 Conformance little change.
Conformant points should read Conformance Points. (2.2). 3. References updated to current versions of all the referred to docs. Non normative updated with latest ISO 10962. Also the DL handbook since we are now using DL for our notation. 4 Terms and Definitions not touched. May need some revision - ditto Abbreviations. Section 6 Additional Information 6.1 unchanged, but Pease review. May need revisions we are focusing more on a technology audience than we were when this material was written. Updated the organization list
Action: Ian -6.3 NEW please review. Explains the DL used to show the restrictions. Slightly modified to be more in line with set theory than DL symbology. Restrictions use the math notation not structure English.
Pete: How do you distinguish universal quantification from individual value? These are the same in DL but are declared differently. So e.g. genders can be declared as classes or as individuals but these mean the samething in DL.
Given we are documenting OWL is this right? No, because there is no notation for OWL specifically.
Pete: if the idea of this table i to unambiguously define what the OWL means, so surely this should be detectable from reading this table? For example, the hasValue restriction is the only place where we distinguish these things. If we use the math to represent the axioms these should be Ok
Complement and Disjoint seem to have the same notation. Complement is used rarely (and not so far in FIBO). So we might just delete Complement.
6.4 Describes the changes in Foundations. Summarized here. Do we need the detailed point by point changes as well? Yes we do. These will also go in Section 6, along with the summary of the changes. We tried about 3 different approaches to how to describe such changes in previous RFCs, and were pushed back each time. This time they will all go in 6.4.8. For example 6.4. gives the new tables etc. already for the new material. This is an example of the new notation usage. The text in bold is the DL notation description. The thing in round brackets is a dummy identifier for the restriction, and these are then listed in another table. Where no qualification is given this applied to "Thing" and is known as an unqualified restriction.
Action: FND FCT need to review Quantities and units see diagrams - this week!
Domain and Range as terms, are not included in the DL table and probably should be. Then see 6.4.4. for an example of the detailed change instructions. Includes addition to the text in Foundations as well as the ontologies. These are at the level of detail needed for "Chapter and Verse" revisions of the FIBO Foundations spec. Reviewers please read these as editing instructions for the published Final 1.0 FIBO
Foundations spec. Table change are done wholesale for convenience.
Products and Service’s has a new property missing definitions (as reflected in the
doc). Diagrams will be laid out slightly different to before, as they were recreated in FND from scratch. Applied to Products and Services diagrams. Minor change in how identifiers are used, e.g. special customer, client and product IDs, whereas before it was just a generic Identifier. That’s the only change from prior material that was in FBC.
Page 35 has an example of a complex restriction. PR shouldn't there be a dot between the isPlayedBy and the class? Reference is the Description Logic Handbook. That is an informative reference. PR: Shoudl be a normative reference if we are relying on it to determine what these statements mean. EK did not want to refer to the whole handbook, which has since been revised. We could refer to specific papers as normative references for these notations. Elisa doesn't have the latest version and so did not want to make it a normative reference. If that is the case we should have a complete normative piece of material on this. May add a note about leaving off the dot when you have parentheses. This in preference to making the book normative. PR: Need to avoid confusion with the parentheses around the reference to the Restriction dummy identifier. How? E.g. use a different kind of bracket, e.g. square brackets. Make it clear in the bracket are used differently. Making the brackets bold is not an option: what looks like a complete statement in bold is actually one where all the pieces are in bold but the requirement is that the brackets themselves in these statements are not bold, for some reason.
Example of Disjoint is given in the table on . To deliver on the name change for OccurrenceKind need to be voted on by the Foundations FTF. OccurrenceKind is in the normative spec in the DateTime ontology that Mark Linehan wrote for FIBO. That ontology was not part of the legacy FIBO and has not been reference in current
FND reviews. Will therefore need resolutions to change those. FND RTF will need to action any changes to DateTime.
Section 6.4 needs to include th document number, so it is clear which document is being edited in the change instructions. Instead of "Changes to foundations" it needs to give the OMG document number. That's the changes so far.
Action: Elisa will work with NoMagic to find a way to output this stuff in Excel, for onward reports because there is no report from MagicDraw with which to generate these things.
The DL descriptions are not included in the content of the ontology in VOM, as these
use ODM not DL notation. If you wanted a report from MagicDraw you would need to add the DL stuff into NoMagic itself as part of the class notation.
Would this include the (set theoretic) symbols? Cardinality maybe easy, but the ForAll symbol and the ThereExists symbol will be a challenge to add in MagicDraw. May be able to write a business rule to do the transformations. Would want this to be rule driven not manually added as annotations to the elements.
Please review. Need the final Foundations ontologies and most of the FBC ontologies in here by the end of this week. Also feedback on the structure, from the business SMEs please. Whether or not this notation works for you so you can review the content. Also whether the narrower tables are easier to understand and consume. Ian's initial reaction is yes, but will study the spec in more detail.
Demo in Protege of some things... Using OWL, shows that given individuals that are parties in role and legal entities [actually Functional Entities and legal entities, not the same thing! MB]. [Functional Business Entity and PartyInRole are siblings and should be disjoint]. Given data bout 2 Citi related entities, with their roles, the reasoner is able to tell you who their primary US Federal regulator is. Also identifies who the precise entity is. Using the US Jurisdiction specific ontology, given where Citibank NA is defined as a National bank, and given that thy were first insured, the date they were established and so on, and other stuff, this lets the reasoner detect a bunch of things about that institution e.g. who is the primary Fed regulator (the OCC), and that it is a member of the Federal reserve system.
Have not integrated David's new material where it would state that a given entity owns and controls another given entity. This would show, for real banks, the equivalent of what David's PoC shows for Globalbank BigBank etc.
Have a handful of banks and registered entities as example data. Has included 11 Registration Authorities from the US, there are a couple more to add. (lists a set of regulators, not registration authorities). As Pete has pointed out, may need to refer to the Joint Venture between TCC and SWIFT for some of the functions
Action: Elisa will then ask Citi and State Street to review the list of Regulators.
David Saul: Just US?
Want to include another Jurisdiction specific ontology for regulation specific definitions. For e.g. Reg O and Reg M (still in the US). Others to add which are not clear yet, and distinct banks for Fed and FSOC and so on, and the Foreign cited admin, some specific states and so on (which would be in a US Statespecific otnoloy if we did them). Some of the above are in the model actually. Will integrate some of them into examples.
Possibly missing: HQ for Citibank NA in Souix Falls SD, would then be able to infer which Fed oversees it. Not yet done. May leave this. No need to worry about that there is more than enough information in these examples already. All in GitHub except examples, and may choose to split out the state jurisdiction piece.
Don't have the equivalent concepts for the EU e.g. having the ECB as an individual of the type Central Bank for the EU. Can add more of these over time. Need to be able to support the very different arrangement in the Eurozone, where NCBs report to the ECB, whereas in the US there are 5 federated Central Bank units but no state level central banks reporting to them. Each Eurozone NCB reports to the ECB but these are not necessarily "members" in the same sense as in the EU.
David Saul: The equivalent for the Fed would be the individual countries Central Bank. For Ireland whereas the ECB represents a looser federation. These are NOT the same concept. May have to create a new kind of class for the kind of thing that the ECB is. May need to represent this with properties that may need to be unique.
Gareth is working on this.
Depends on whether we are trying to model analogous data, or capture the actual legal business meanings of things. For example what is the formal meaning of "member" in the target data. The goal was not to have a crisp definition of the meanings for the Fed but to demonstrate that things could be represented. FBC itself is only guidance, not the formal semantics of things that may be needed in jurisdiction specific ontologies. MB suggests we would need to identify what concepts are applicable, so these can be stood up in e.g. Eurozone, China, anywhere else. So e.g. we have a concept for member banks in the US content but we have not
attempted to stand up similar but semantically distinct concepts from different jurisdictions. We would not stand up additional concepts from non US jurisdiction without a use case to drive these. Per Elisa.
What does it mean to have jurisdiction, for the Federal Reserve? This is the geopolitical area over which a given district bank of the Fed (e.g. San Francisco) has jurisdiction over San Francisco
PR: What does it mean to have "Jurisdiction” in this instance? EK: this is relates to what that part of the Fed is able to do. Does this mean it can regulate companies doing business in that jurisdiction, or other things that governments can do? No. Relates specifically to the jurisdiction which the branch of the Fed has in respect specifically of the laws under which they administer these banks,
Is there some document we are not bothering to model that says what exactly does a Federal Reserve bank regulate? Yes. There are many such documents. DS: This is correct. So e.g. the Boston Fed has jurisdiction over things in its region, e.g. if they have a determination of facts against State Street, this this comes under the jurisdiction of that branch of the Fed. This is because the bank was originally chartered in Massachusetts. Regardless of where the company is registered, it relates to where they are chartered. However, State Street is both registered and registered as an entity in the same state so that question doesn't arise, as it would for Citi. Tee bank charter states where the bank is headquartered for the purposes of the given litigation, which answers the question of under the jurisdiction of what Fed does the bank fall. E.g. State St is still a Mass chartered bank. And as such, a member of the Federal Reserve, at least for the banking part of it. The Federal Reserve also has banking jurisdiction over what are determined to be Systemically Important FIs, so it's not impossible that the jurisdictional reach for that set of regulations might be elsewhere, e.g. New York state and Fed. The legal entity in the case of Citi is in De but the bank is HQ'd in South Dakota, but was originally chartered in New york City. so their charter was revised to give the HQ as South Dakota.
Action: David Blaskowsly and Elisa to work together to create individuals for State Street as well, for examples?
- Elisa Kendall Determine how best to incorporate the examples in the body of the spec from public information about the banks - FBC-23Getting issue details... STATUS .
- Mike Bennett Mike B to review Section 0 edited. 0.3 submission team updated. 0.4 minor language tweaks. The basis of this text is IND. MB also to review 0.4.2 - FBC-24Getting issue details... STATUS
- Pete Rivett Pete to review Section 2 Conformance little change - FBC-25Getting issue details... STATUS .
- Ian Maung 6.3 NEW please review. Explains the DL used to show the restrictions. Slightly modified to be more in line with set theory than DL symbology. Restrictions use the math notation not structure English - FBC-26Getting issue details... STATUS .
- Mike Bennett FND FCT need to review Quantities and units see diagrams this week! - FND-25Getting issue details... STATUS
- Elisa Kendall Work with NoMagic to find a way to output this stuff in Excel, for onward reports because there is no report from MagicDraw with which to generate these things. - FBC-29Getting issue details... STATUS
- Elisa Kendall Ask Citi and State Street to review the list of Regulators. - FBC-27Getting issue details... STATUS
- Unknown User (dmblaszkowsky) David Blaskowsly and Elisa to work together to create individuals for State Street as well, for examples?
FBC-28Getting issue details...