2018-02-13 Meeting notes
- Dennis Wisnosky
Date
Attendees
Agenda
1) Use Case reminder
2) Where we are on our road map.
3) Open Action Items
4) JIRA Issues Review - https://jira.edmcouncil.org/projects/FCP/issues/FCP-4?filter=allopenissues
5) Todays content discussion.
SMIF OWL-UML
SKOS
RDF/S
6) For next week.
Proceedings:
DN changing graph approach to not set up annotations to exclude properties. Because this then requires annotations to include certain concepts. DN approach will include everything by default. Then will only intentenaily exclude some concepts. PR to DN Don't we wan to use SHAKL for this? DN in time yes.
Elisa has done some work, but has not issued a Pull Request. Many issues done, not closed. DN to EK, just close them. DK many definitions complete. PR has a checking tool that has helped EK on DER. EK to PR, don't run yet.
EK to DN Want to eliminate all properties that do not have labels. DN - Do it.
JN I there a way to show the IRS ontology alone all by itself so that it is easier to understand. DN Yes, but there are so many related concepts. We could eliminate those concepts extraneous to IRS. DN What is meant by show. JN can we just show those concepts in Protege that are pertinent to IRS. DN is showing that now on the screen. Is showing only what is related to IRS. There is an About file that does this.
JN in CFTC there will be certain people who only care about IRS. Nice to have different views as a selling point. DN Yes, that is the purpose of the about files. Because we have not built out other DER concepts, this is focused on IRS. Could build About files for Credit Default Swaps and others. Could do this on demand. JN makes perfect sense.
DN has approval to make his presentation to CFTC as long as it does not have Wells content. Can show IRS concepts and taxonomies. Can't show relationships between elements in the ontology and the source data because it connects to Wells confidential data. JN Understand, just say it, even can't do it. JN need to get off line a orchestrate the presentation.
DN want to be able to cover all of the concepts in the spreadsheets. DA needs some more coding to show all existing relationships. EK Yes, they are there. DN can show unions and more. Another week or 2 and we will have it nailed.
DN to EK Looks like you are pretty much there. EK, just a couple of issues to discuss with DN. Small things. Last week did webex with OFR. Excited about LEI Level 1 data. Gap now is that relationship information is not modeled in Business Entities. All of the info in the CFTC spreadsheet will be relatable because of this LEI work. Will know who is related to who. DN Really looking to changes in BE. Make it more simple. Eliminate abstractions such as legallyInforporatedPartnership. Does not exist in any known jurisdiction. PR has 22Million LEI Level 1 data triples. Can demonstrate for example if an entity is an owner or not without having to declare. All is in GitHub in Release. Not het in production. David Saul has been doing this for State Street. Need to be reviewed by Wells. PR has take a more mean and lean approach to creating triples. In an operational system do not need a label on every LEI. Won't scale with that. EK meeting with OFR again about LEI and NIC relationships and unique loan identifier. Jeff Did similar thing about a year ago.
DN wold like to see the changes in BE. Would like to see these relationships in data graph. EK and PR now working on ownership and control relations in BE. Need to run past Jeff.
DN Because we are essentially done with this DER work let's use this time to shift from IRS to BE discussion in this time? Will help Wells and be additional source for the data graph. Should also include the individuals to ID reference data. Expose the currency codes.
EK FND has much of this. PR has done some heroic work. DN would be good to expose this in the data graph. The data graph is simplifying the ontology for more operational use and eliminating the need for inference processing to actually see the obvious in the data through the graph. Much more efficient to visualize. Then can link easily to individuals to show relationships. Can show all of the members, e.g. of currency. DN to JN would this be attractive to CFTC. JN Yes, very much so. Very powerful message.
DN to EK Next week can we shift gears to BE and LEI? EK need to talk with PR about some concerns on Options. Do not what to switch over entirely. DN Yes, maintain focus on CFTC. Work on LEI within BE very important to CFTC. Must also include the counterpart info. PR can show the LEI work. DN Want to see this in this in the data graph. The 'molecular structure' of an IRS. JN Yes, this would be fantastic. Spend lots of time doing work brute force.
PR is there documentation for doing the annotations for data graph, or is this all in Deans head. DN it is in the code in GitHub. Would be part of the FIBO release. Another version of the RDF/OWL.
PR The IRS data is based on accounting roll up. Who understands how this relates to ownership and control? Pete Maybe Jeff or Bruce. Jeff Basic idea is 51% ownership or more drives accounting consolidation. Not one of ownership and control, one of accounting consolidation based on percent of ownership. EK will need to ask questions later.
Decisions:
Action items
David Newman Send GitHub link to Deans code to Pete.
David Newman Dennis Wisnosky John Nowlin Discuss CFTC presentation.