2017-08-03 Meeting notes
- Dennis Wisnosky
Date
Attendees
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.
Jacobus JIRA
Randy FYN errors
[Future: strongly recommend we provide, in addition to the FIBO-specific flavor of SKOS above, a more conventional SKOS vocabulary in which properties are rendered as concepts. Also suggest that in both flavors we enhance the script so as to make our FIBO-specific annotation of “synonym” become the better-supported RDFS alternative label annotation.]. MB
5) For next week.
Proceedings:
20170803 FIBO FPR RDF ToolKit
Focus on Action Items. DA has identified the ones he has done. The About to All change has no yet been implemented. DW and JL OWL Slide words to be written - needs them to talk. EAL not seen in a while
JG Cambridge assignment - Not completed yet. JG has been working on this, did not pass initial test. Expects to finish this weekend. JG did see one other thing on this - the new directory listing generator. In this you don't see the differences between versions you only see the Master version. So we are missing one step since you cannot select other versions .
PR there are still broken links in these -see correspondence with RC and PR on FYN errors.
Need to publish Production quarterly, whereas Development which we see here (i.e. Master) is continually updated during the Quarter. At this point, until the JG script is complete, we can't output to the structure currently in spec……. Solution is to have one page before this page, where you can select the version.
JG will finish the script at the weekend. We will work within the current web structure while we fix those things. The long term track will implement the requirements from Pete's OMG group (one IRI goes to one page and you select products at that level). This will be implemented within the new HTML architecture.
Next action on JL OWL Slide Will be worked on when people come back.
Diagrams rename for (AB) - MB will do today.
Naming convention for IRIs and labels from ISO 20022 (MB) - to do
MB to work through the results of SP10 testing and identify mismatches between CCM and OWL that are impeding round tripping. MB to do (top priority today)
OK action on Fragments - was done but did not go in the June release. will go in September.
PR to send JG link to React component action. PR will re-send, may have done but JG not finding it.
JG screen: alternative to OWLDoc documentation generator. This is called OntoDoc, used for their internal ontologies. Demo on screen. Lists the OPs, DTP, Named Individuals. Also has a bobbly diagram auto generation feature. This is incomplete as they have not yet got all the right labels. The demo includes devices etc. tech stuff, and also activities, agents and so on from Prov-O. PR: What is the audience? JG Not someone who wants to validate the ontology but someone who wants to browse and discover. PR see list of possible audiences in their group's work - need to identify which audience this is for. So the viz also covers external ontologies that are referenced.
MB: Looks like a specification so it could potentially be used to generate the engineer-facing formal documentation as part of the OMG spec (not for a business domain audience). DW: have they tried uploading FIBO into this and see what is looks like? Seemingly not yet. JG recommends we do that. Put it in front of people and see what they think, what audiences it is a fit for. DW how much work to make that happen?? Do we import FIBO into some tool? where does it reside? JG it is a command line tool, you can run it over a directory of TTL or RDF/XML OWL files. Generates a bunch of HTML files including the visualizations.
PR we have the diagrams in SMIF.
JG more of a gimmick at this point. What is the effect of modularization? You get one page for one OWL file. Can link onwards to other OWL files in the graph. So a modular design is easier to explore. MB: this provides something for a tech audience whereas UML does something for a business audience. Both equally important but not the same.
PR the user reading of the SMIF diagrams needs to be better documented. PR: We also need to better document the relationship between the SMIF constructs and OWL constructs for an OWL audience. JG Modern tools are consistent in every visualization. Possible to go from one to any other without loss of meaning. DW Yes, that is our goal.
PR: 4 views of any ontology: Glossary, Taxonomy, Structure and Ontology. These are the 4 view types. There are also the audiences, which is not quite the same. e.g. structured view can be done for business facing and for tech audience with different things.
Next topic: DW: for the FYN, what should we do about broken links Randy has found. FAO-Country is not used and can be removed. W3C time is also snapshot and used only in Ext. There were things PR could not get to from the tree navigation and the URI resolution didn't work e.g. when putting the namespace URI in the browser. Something odd both for tree and namespace resolution.
JG to MB, maybe it would interest you to think about what the semantic meaning of the position of certain elements in a (UML or any other) diagram is? Isn't the fact that you position one class in a prominent position (for instance) interesting information that we should perhaps capture as a special annotation or so? If we could do THAT, then we can actually generate diagrams that users would like...
RC from a user perspective the most important thing is the FYN from the About file. You can come in from a lot of different places. If you don't have the insight. The double slash isn't necessarily a stopper but it tells me that something that went wrong.
DW do we have a list of all these breaks? No, these come up based on the context RC is looking at. RC this morning's ones are individuals for e.g. European financial services individuals. Does JG have any insight on why we would be hitting these, e.g. the 403 resolution errors? JG needs to understand what the problem is. MB suggests that individuals are different enough that they might be giving us a systemic problem? to PR screen...
AccountingExt under FND. This has double slash. This suggests that something has got appended to something that it shouldn't be. However the URI looks OK once the double slash is removed. This one gives 403 (forbidden) not 404. JG has a theory about this one. Also 404 not found in BE. JG This is a change in nGnix rules, it seems to be handling these in some different way. Experiment with what the IRI should be. Then it works. so there is some issue in the nGnix rules. There were changes JG tried to merge from Mirek Sopek for the nGnix routing rules. That may have messed something up. Here the IRIs are all /master/latest/
RC this suggests there would be errors throughout. MB to remove FAO Country. MB to change refs to W3C time to be live refs to the actual ontology. Then the /ext folder structure goes away. JG knows what to do on nGnix. JG will do that change, and Mirek will need to do something to compensate.
Next action: SKOS Deliverable. MB outlines the 2 flavors of SKOS and suggests we at least describe why we do differently what is different. PR also what we put out in SKOS includes a whole lot of owl stuff. This may be a result of this FIBO flavor, but agree that a generic SKOS implementation would not support this. PR agrees with MB. JG agrees with MB.
DW So this is another Product for Sept? Yes.
How to generate this? DA would need to create a branch of his SKOS creation code, which include much of what he already had. JG we can also monitor who uses the various products and in future deprecate ones people are not using.
DW - this brings us to the testing that has to happen when someone asks for a pull request to be honored. Also at present DW assumes that something that was tested in someone's branch will work in Master. This must be a rigorous test, not a hope.
DW how do we track who is using what? How do we get feedback like we are getting from Randy? We were counting on the new EDM Council website on this but that is slipping into the future.
PR we should start to create a plan working back from Sept 30 so we know what has to be done when, and have a cut-off date of a week before. This would be the end of the week before OMG.