2017-11-21 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:
JN: I understand that we can go directly from the conceptual model to the triple store. We don' need logical and physical models. DN: Yes!
Dean still in class today.
DW hours Elisa pull request that eliminates erroneous restrictions.
DN shows the Allegrograph FIBO diagram for IRSs.
JN: We are still waking for OGC to give go ahead for use of anonymized data. But we can prove our points with just the Data Dictionary. And, we can show the diagram. The Data Dictionary does not need to be complete. We just need to show the connections. Today these are separate efforts. We can show that this is one effort. All are created from the ontology. These are not separate and do not need human intervention for manual updates. No stove pipes. We can demonstrate and show that these are all connected.
DN In Protege we can show the actual logic. FIBO has both operational elements and conceptual elements. Need to enhance FIBO to make it more clear what is operational and conceptual. DN to JN, What do we need to make the existing model complete. JN: Need to discuss with Robert. People at CFTC don't know what an ontology is. Do know conceptual models and DD's. We don't need all 88 data elements. Just need to get buy in on the concept of FIBO and get leadership to agree to invest. We can just show the vanilla interest rate swap. That is enough. They will buy inform that. Don't need complicated swaps. Just show that the ontology generates all of the components that CFTC leadership already understands. We put together a presentation and show JN and then set a date.
DN We could further improve, but if JN thinks what he sees is sufficient, then we can go forward with this. JN might get comments from Robert, but believes what we now have is sufficient. Just need to put together a presentation that flows. DW to DN - You put together the absolutely best presentations. DN: Ok, but need to get with Dean one more time to thoroughly flex out the logic. E.G. Child inherent from parent and therefore should be bigger. Just need to resolve some of this logic. DA has all of the code so he is responsible.
ACTION: Dean to check in all of this work to GitHub.
DN to JN: Can we do this in early December. Yes, but want to see the presentation before sets a date.
DN to EK: Are there any gaps that still should be plugged. EK Some SEC work is relevant to this re product types. But, does not need to be done for a demo. There are some definitions that could be enhanced for reporting. Not important gaps for instruments like DN showed today. Content around instruments is there. State Street can give us more. But, we have enough now to prove the points that JN wants. Does not need to wait for more of these concepts to be finalized.
DN: The story line is that by leveraging the ontology we are taking a big leap beyond relational data bases where every component is independent and all gets out of sync. With an ontology, this does not happen. And, where there are similar issues, automation can fix them. FIBO mitigates risk to a manageable level. We have great presentation material and just need to work on a specific story line. Could show queries such as around instrument hierarchies. Can so the graph and how simple it is to work with the graph and the underlying data.
JN: Could start out with the diagram that DN displays in allegrogtraph. Explain where it came from. Then, show the DD and that FIBO also generated. Show the actual FIBO triples. Make the ontology real by showing how it is the source of all products.
DN: Good! We can show many queries over the data and demonstrate what is possible. JN Everybody will understand the diagram. The audience may not understand the benefit of queries. The audience will not be people who write queries. DN: The quarries are written one time and can be used over and over again. Could not use Allegograph Gruff. Could have a more simple way of showing graph visualization. Could show more behind the graph. JN: That would be fantastic.
DN: We can show examples and then CFTC can take it from there with coders in DC. JN: Very powerful to show what Wells Fargo is doing as representative of the direction the industry is going. DN: Wells can make an investment and demonstrate the power of FBIO through Wells developed code. JN: Yes! Show the examples and then the actual ontology. This will sell the entire benefit.
DN: Just need to fix a few things with Dean and fix a few things.
EK stop focusing on more and instead focus on tight crisp anticipating questions that will be asked. JN present the team and why, then 1 slide on who the Council is and then DN a hand full of slides on what CFTC wants and needs and how FIBO answers their questions. One slide that show each answer and that each came from FBIO and then show the ontology at the end. Then, we do the ask. and how the end result will add big time value to CFTC. Fifteen slides max with ability to go into demo at any moment when a question is asked. This is a DARPA type of presentation that works in Government. Must be reversed and tight.
DN: Taking to a Finance professor in Santa Clara who is connected to other professors interested in systemic risk. Could we get the anomomized data and share with him. He is doing research that may offer insights on risk that could not be realized in any other way. JN: When we get the go ahead, the data should be releasable for many purposes. DN the professor knows Shrini of CFTC.
Will work on the story board next week.
Decisions:
Action items
Dean Allemang check In all of his work to GitHub