2017-06-22 Meeting notes
- Dennis Wisnosky
Date
Attendees
- Randy Coleman (Unlicensed)
- Jim Logan (Unlicensed)
- Mike Bennett
- Pete Rivett
- Anthony Coates
- Omar Khan (Unlicensed)
Agenda
1) Where we are on our road map.
2) Open Action Items
3) JIRA Issues Review - https://jira.edmcouncil.org/projects/RDFKIT/issues
4) Todays content discussion.
WHERE R we and who needs to do what for delivery to Dean on 0700 US EDT 27 June, 2017
Spec landing page will have:
Glossary: goes to: HTML glossary generated by CCM
OWLDoc Generated by Protege
Vocabulary: Goes to SKOS
SMIF: Goes to UML and matching OWL
Ontology: Goes to many places
index
NQuads
.ttl
etc. etc.
ALL of the above goes is in both Dev and Production
5) For next week.
Proceedings: 
20170622 FIBO FPT RDF ToolKit
30 June Deliverables
DA Main thing the round tripping of OWL into MagicDraw. Last night, took a list of URIs from MB that have been adjusted in CCM FIBO-VFull, so that the OWL can import provided it has the same IRIs, which DA has changed so that it does. Is doing this in FIBO-Master. Today, would like to identify when we are ready to do the switch.
MB will run the Cory macro today, once as test and then as real thing, and then we are ready. Meanwhile MB sent an email a minute ago about the possible collisions of annotation properties. JL if we merge the project we will have both sets of annotations. Need to unify the annotation properties. Otherwise you only get definitions for one of Pink-FIBO or FIBO-VFull but not both. JL: In MD 18.5 it has addressed a shortcoming in the refactor action (ability to refactor the use of one of the SKOS copy with the other copy). This works for classes but not properties in 18.4. We think this is fixed in 18.5. We also need to continue to coordinate all products and profiles across all FIBO users, while we are doing this. There would be about 5 annotation properties. If that does not work, then we would need Nilesh to write a script that would iterate through these and find the annotation property in the other product that has the same URI.
MB now seeing UML Documentation even in the absence of a setting in FIBO-VFull. JL the reason for this is that setting that option is not reversed when you remove the setting - it is sticky. What is happening is that the Project Option isn't sticking. So it got itself unset, but the individual annotations are still set. So they are still not the right ones, so the problem described above still exists. The same nominal annotation is selected in both projects. By nominal we mean the same IRIs. However, they have different GUIDs. Version 18.5 may fix this. May be available ahead of our Jun 30 date. JL can try this on a branch. If that works no-one else needs to upgrade to 18.5. We only want to move everyone else onto 18.5 as a whole, after June 30.
MB any content changes before our drop dead date? DW Yes, we now plan to make changes in Loans and the corresponding changes in FND and FBC that MU, EK and MB are corresponding about today. MB if we make content changes in Pink FND then the way use datatype properties differently in the 2 projects may cause interesting coordinating issues.
Meanwhile the JL thing that is available by and of day tomorrow - is in US time. Does either the script or the 18.5 stuff require working on the 2 files individually or on the merged file? JL - the end result is I should be able to edit FIBO-VFull and Pink-FIBO as they are now, eliminate the duplicate APs now, and then it will be ready to be imported into the FIBO-Master project. Can we schedule a meeting late tomorrow to execute the merging (along with the Proxy script)? Will ingesting OWL make the proxies disappear? JL the OWL ontology is compared with what is in CCM. If the incoming ontology is missing a class, we make a deletion happen.
Friday June 23 US EDT in a GTM session: We create the unified Master, do a test import of OWL to make sure it all works, then MB moves the blunt ends. Then Monday first thing ET, DA and MB finish importing the OWL. Change to this: MB will move the tails tomorrow in FIBO-VFull rather than in Master, so we have a de-proxified FIBO-VFull ready to do, and the Pink and the OWL. Also implement the Loans namespace (Loans Upper Temp) per recent correspondence. This is what JL will be expecting from us.
FIBO-VFull = DirtyPink + Extensions + future. Pink-FIBO = Pink. Both together = master (in Git). Unified CCM project FIBO-Master corresponds to Git master when this is done.
Next: Status of OWLDoc Command Line process (TC) TC: The idea was to produce OWL documentation. There is a plugin for Protege that does this but we don't want to run Protege we want to run on CL, The guy who produced the plugin said it is possible to run this from the CL. Has sent a simple sounding description of how to do this, and TC has been following that. TC set up in our Build environment and it did not work. Then created a new environment with none of our other stuff. This works but, at the end of running, it calls system exit which kills any Java process inside which it is running . This is not good for our toolkit generally. TC is looking for a way to trap that call to system exit. According to info on the Web you can but in practice it is not so easy. TC also checked out the source code to see if he can add an option to the CL to not call system exit, and do a pull request for them to create a new version with this feature. But TC can't get hold of this. TC also don't know how long it will take to resolve this. TC would like to know to what extent we can engage Matthew horridge, get him to add that option to stop the system exit being called. DA there is an issue with that: Matthew horridge is not the person who built the plug-in. That's the person running the guest house on Skye. DA even when it runs in side Protege 5.2, it actually makes a mistake in the OWLDoc that is creates. The mistake being that the frames are not wired up correctly. There is a missing target thing. So suggests that TC crack open the source code anyway, and fix that bug. TC to make it work you need to tell it where Protege is installed as uses some of the JARs from Protege. But one of the JARs it says it needs does not exist in Protege 5.2
What about other tools? https://github.com/dgarijo/Widoco/ PR: Elie found a this. This gives a console command line the use of which is documented. Have we seen its output? See Gallery on that link. DA can run this on FIBO and see if it satisfies DN requirements. PR: this is built on LODE
MB there is also OntoGraph (Nine Points) to look at if we are still stuck, but that is also being refactored at the moment. DW: Are we to give up on OWLDoc or continue these things? DA we do not have a workaround for this. So we do not have a safety net. Can someone try out Widoco as our new safety net?
OK: Took a look at Widoco a couple of months ago but could not (via the UI) get it to run. Might have been an environment issue. DA this might not be an issue if we are using it via command line. Who has time to look at this? TC can look at Widoco and see if he can it to run.
OK Progress: Going well, JG opened up the ports, OK using the fragment server to pick up the various versions, get the versioning working. No issues with that.
DA: That is everything I needed.
TC needs JG to get the Build environment working again so TC can push out his fixes to the Master branch so that it works again. TC has given JG a JIRA issue ot action this but had no feedback on when he will find the time. See JIRA RDFKIT-80.
DA: For FIBO-V do we expect 2 publications, one for Prod and one for Development? DW thinks we said Yes but we should decide this collectively. Consensus yes! DA agrees.