2015-10-19 Meeting notes

Date

19 Oct 2015

Attendees

Decisions

Proceedings

Presentation Slides –

20151019 FIBO-FBC FCT

Not yet made progress in incorporating Gareth and Tracy's comments. Receive no comments since we published the RFC. If there were any comments, we should get them on the EDM Council wiki so we can work them off. That was the main homework until today. So assume everyone loves the spec as is! Gareth has a comment pending, which needs to be revised in light of the most recent version. some of his comments may have been addressed but not all will have been. Also some of Gareth's questions will be in JIRA issues that we have not got to. Gareth to please add a comment to the JIRA issue to confirm it is still outstanding, or to move things along. Gareth wanted to be sure the time was right to raise these, and will walk through some with Elisa to vet that these are sensible issues. Can then discuss these with Elisa, possibly helped by Shaun Rolls and Fabien.

Slide 4 scope as previous. Note EU ones planned. For international ones GLEIF Registry would define content is that regarded as International, or EU and US? Assume International. But is International really a jurisdiction. Or should we just drop it in the registries folder. Clarification: All of our individuals so far ahs been US specific. There are more to create e.g. FINRA, self-regulating entities. For some of the exchanges, markets etc., some will be in US or EU or e.g. the UK (which is also in the EU) and so on. Elisa needs to know which exchanges are relevant to people. MB notes that in the securities markets there is not a matter of some being important and some not if are in the international securities markets, this isn't walled off like jurisdiction specific retail products and the like. What's important is the concepts of market, exchange, MTF and so on, not having a small list of half a dozen exchanges. However, exemplars for a few of these make sense. Exchanges may also come and go at a higher frequency / volatility than market regulatory infrastructure. so the emerging consensus on exchange and markets seems to be that we need exemplars of the main ones.

What would be good exemplars? what ones do we have as exemplars at present? As distinct from US Agencies which are solid parts o FIBO not exemplars. Note that the use of the word jurisdiction here may be slightly hazy for example the UK and the EU are not mutually exclusive places the UK is part of the EU. So maybe Elisa means geographies of some kind and not jurisdictions? Also the UK is not a jurisdiction at all. Considering where to put exemplars waiting on some guidance from Dean or David on this. Also waiting on someone to action the recent pull request in which these new exemplars were uploaded.

FBC/FunctionalEntities/US Jurisdiction is where the exemplars are that were just uploaded. This is where there exemplar (not normative) bank examples such as Citi. Pinnacle Bank has also been included das an exemplar of a small regional (and statespecific) bank. California is both a state and a jurisdiction. We will need jurisdiction instances for every US state eventually. The registration of the banks is distinct from what you register a business with for a state bank these are registered under and regulated by both agencies, in CA. Delaware is widely used for business to register so this has been included. >80% of banks and >40% of companies in the US are registered (and registered in the other sense) in DE. Apart from that all there seems to be there is a service area on the freeway :) Elisa gives a walkthrough of the exemplar ontologies, via a text editor. In the Citi example, one of the kinds of registration that as a business, is in EU, while the other that of Citi as a bank, is in Sioux Falls South Dakota. This provides a kind of template people can follow to create their own example individuals. A good challenge here would be in considering how to model Deutsche Bank (note that Germany is also a federation).

Q: Should the exemplars also be in the FBC namespace? Yes, these are examples and are created as such in the FBC namespace. Others can create their own exemplars in their own namespaces. (I presume that's what Elisa means). Q: Isn't this as much an example of BE as well as FBC? Yes, it is but with more detail that would not be available in BE. Some of the classes that are instantiated here are in BE. So in answer to the above question should the URI for the individuals be in FBC? You can put them in BE if they did not include concepts in BC, such as business registries. The class Jurisdiction is defined in Foundations but the concepts referring to this are in FBC. Clarification: if you create individuals you can create them in your own e.g. db namespace. There is no compelling technical reason for where the namespace is, but the exemplars are published in FBC because we are publishing them as part of the FBC exercise. The reason Gareth asks is to be clear when one is using something that is core to the specification and when you are using something that is not. So should all examples be prefixed e.g. example individuals. That needs to get fed back to the other teams so they use the same approach to representing exemplar information. Similarly the jurisdiction specific folders.

Elisa has talked to David and Dean who are apparently talking about creating a single folder for jurisdiction specific stuff. The question there is, would we have one folder for "Jurisdiction specific" stuff, with jurisdiction specific ontologies within that. However that arrangement, while it might work for partnership types, is not going to be future proof for things like FBC where we might need more than one OWL file for each jurisdiction or geography. So this needs to be discussed on the Leadership Team so that we have one strategy across all FIBO,s that would be big enough to support all the likely requirements. ACTION: discuss jurisdiction exemplar strategy on FLT.

Gareth is having trouble loading this stuff into Protege without it falling over. This seems to relate to examples a lot of individuals will cause the Hermit reasoner to fall over. Dean needs to talk with Ian Horrocks about this. Dean is also running additional tests. Meanwhile Pellet works users need the pellet .jar files for Protege. Gareth has this so it's not the Hermit usage that causes this problem. There is something else happening. Gareth and Elisa to liaise on this . reminder: in terms of the namespace strategy, we still need to identify where the LEI / GLEIF stuff needs to go. Possible an International region (it is not a jurisdiction). FYI all of this stuff is current and there are no changes in GitHub from what we submitted the only exception is the Business Centers stuff, which can be added via Securities if needed. How quickly would we expect FIBO in OMG to get updated. For instance FpML has a regular update cycle. So, from the time we make a change, how soon does it become part of the canonical FIBO? See LCC for a good example of this question. This reflects the currency and other codes. These are updated periodically by ISO via a spreadsheet. So in OMG, the Revision TF would have the role of monitoring changes in the ISO specs that the LCC work is based on. Each time there is any material change, would raise an OMG JIRA issue against the RTF. This can be done in one meeting cycle (3 months) or synchronized with when there is other material to modify the spec with. These are two entirely separate cadences the speed with which a code list in ISO is published, and the speed with which industry changes as reflected in the kind of cadence we see for changes in FpML. So these are 2 questions not one. The pace of the business merits a different approach than what the OMG is generally used to when they deal with modeling standards. This has been understood (or at least, the EDM Council has tried to convey it to folks at the OMG) right from the day we joined.

Slide 7 - there are some definitions we need to tweak. The definition for "Financial Instrument" is still under construction. If anyone has time this week to look at that and comment, please do so and comment in the JIRA issue if possible. This is JIRA #59. See also #61 and #64, and #60 for Central Bank. Meanwhile the proposed definition does not seem to be in the comments on#59? There were emails. This was after we replaced the ISO 20022 / ISO 10962 definition with something broader. (that is, our concept is broader than ISO 20022 / 10962) and therefore needed a definition that reflected the new, broader scope. These are filed as "minor' in JIRA. Any of the "Minor" but non technical issues please investigate this week a shomework. That is, all the definitions. ACTION: members to review the issues with definition in their headings, this week. Even just ideas as starting points for a discussion on definitions next week.

Elisa will also have new definitions for e.g. self regulating organizations and soon, during this time scale. Also definitions of trade, position, holding, exposure, some of which are in legacy FIBO also, which Elisa will use as a starting point for discussions in Nordea and with this group, on definitions.

Q: Status of SEC activities? Plan is to go through as many of the issues in FBC that would impact Securities as possible, by mid November. Anything needed by the early November 4 week date? Just MB action on redline spec, and an action by Pete and Elisa on content negotiation. What about Specification Metadata? That one now works. See page under INFRA from Pete on what should happen through content negotiation for various kinds of requess. We need to test this for each of the use cases in that SM is now resolvable. The version is still the March 14 version. See page on INFRA. There are 3 use cases. There are some alternatives we need to choose between for these 3 uses cases. This is in the INFRA Wiki Pete Rivett https://wiki.edmcouncil.org/display/INFRA/Content+Negotiation+for+FIBO+Ontogies.

Q: did Elisa test the version specific as well as version independent URIs? No, only the version independent ones. Joe Harth has been helping us with this content negotiation business. This is subsequent to ou discussion in the Monthly Update Call where we heard a few user scenarios for what they wanted to see on this. Tested for Contract in FIBO Foundations. One issue is that the MIME type is coming back as text/plain (seen via an http requestor). This is not the required MIME type. One of the outstanding questions is what should happen if there is an accept of HTML in the header of the request. Should we return the MIME type of RDF? Should be should be rdf+xml. See link. This page has not been shown to Joe Harth yet. First we need to agree among ourselves what we want, before we get Joe to do things. See alternatives in this page. For instance XML with accept of XML in the request, should we make it RDF or just XML? What i they ask for Turtle, should we 404 or return RDF regardless. We need input from Dean, Jacobus etc. Also we should do the same on our namespace as on the OMG one. MIME type for Turtle is probably TTL or TTL+RDF or something. Manchester syntax does not have a MIME type. However, Manchester is not a standard format so we should not care about it. We should care about the standard formats. Also XMI, which has its own MIME type. So we should do that too. The XMI is delivered as part of the standard. Then, how do we distinguish ODM XMI from UML XMI ,as these have the same MIME type. It is possible to create a new MIME type, and this could be done for ODM XMI. You would need to apply to the IANA for the creation of a new MIME type. For some reason, XMI itself when this process was carried out, was wrongly put under the Vendor tree (this would mean something in the light of IANA MIME types registration). That needs to be fixed. So now we need to get all the MIME types correct and have the negotiations on the right thing. Meanwhile, we at least have the ability to load FIBO into Protege, which is new. The error Elisa got just now was on Specification Metadata not FIBO. This was missing part of the IRI.

AoB? Also the access file for the SM entry doesn't look right to Pete or to Elisa. How did it fail? Did it fail to get back the RDF file? No, it got some error. It works after all, but intermittently. So whatever Elisa just did, MD is still the one that people are going to have trouble doing. Need to go to File / open from URL. Still intermittent, may relate to how often you restart Protégé. Q: What is the right namespace for SKOS? This is SKOS/core with no whatsit at the end of it. Imports SKOS:core and reference dc properly without extra help or interaction. So once you load SM you get everything we extended from Dublin Core. May be some tweaking to do now that this works. Note that SKOS comes back as text/html when you make a request to it. Maybe Protege doesn't use the MIME type. Other tools will definitely use the MIME type which is why we need to support that. What came back from SKOS when Pete just tested it, was an HTML page, which might be the actual spec or something. Pete tests this by changing the header on the request (whatever that means), to see what happens this time. Changes the requires to RF/XML he gets the RDF back. So you need to do that; by default if you don't do that, it gives you an HTML document back. Which is curiously similar to what Dave Frankel was asking for with FIBO at the Monthly update Call.

Action items

Elisa Kendall Discuss jurisdiction exemplar strategy on FLT.
FIBO Process Team to resolve content negotiation issues.