Date
Call-in Information
Time: 11:00 am, Eastern Time (New York, GMT-05:00)
To join the online meeting:
- Go to: https://duraspace.zoom.us/j/823948749
- Or iPhone one-tap :
- US: +14086380968,,823948749# or +16468769923,,823948749#
- Or Telephone:
- Dial(for higher quality, dial a number based on your current location):
- US: +1 408 638 0968 or +1 646 876 9923 or +1 669 900 6833
- Meeting ID: 823 948 749
- International numbers available: https://duraspace.zoom.us/zoomconference?m=Qy8de-kt6W4fMMDQCAV_3qfH1W-lxAo5
Slack
- https://vivo-project.slack.com
- Self-register at: https://goo.gl/forms/JxQFkut4TYj4Ehww1
- Self-register at: https://goo.gl/forms/JxQFkut4TYj4Ehww1
Attendees
Indicating note-taker
Agenda
Comments? Architectural fly-in updates
Mailing list messages
- jquery.scrollTo error - Do we need to upgrade `jquery_plugins`?
- No Subject - "How can I add more associated profiles for that particular editor."
- Freemarker Template Error
- [vivo-tech] help regarding the custamization
- [vivo-tech] Inferencing engine not adding triples needed
- CODE4LIB list: rdf and doi's - thanks Marijane White
Received
Benjamin Gross : Pending response to code review
(re-)Raises interest in reconsidering first-time, every-time, tdbconfig design
- Should be low-hanging
Where does this stand? What is needed to add more person identifiers to VIVO?
- Mike Conlon : thoughts on where this stands?
Status of In-Review tickets
Kitio Fofack ? Benjamin Gross ? Orcid and i18n
Only touches one file
Low-hanging - need one more reviewer
- Is this feature of broader interest?
- Andrew Woods to look into
- Mostly trivial, with conversation around Tomcat version support
- Relatively straight-forward bug fix
- An important step for i18n... resolves many other open issues
- Low-hanging, documentation
- Kitio Fofack to review?
- Low-hanging... need one more review
Bugs (1.11)
Notes
Comments? Architectural fly-in updates
Aligning ingest and interface concerns with the current work
Be helpful to revisit Brian’s diagram at beginning of this fly-in prep meeting
Combine: Stanford’s work into RIALTO
Having indexing/reasoning etc. be separate components that can be configured/set up as desired
Currently indexing is done in a background thread triggered on edit events
Not to say those threads are lower priority
Since on the same processor, may slow things down
2/3rds of the way there as far as being able to trigger reindexing based on events separately
Bottleneck is still triplestore since index building depends on Sparql queries and does not depend on search index being externalized
Weill Cornell was using Virtuoso but perhaps they are no longer using that
Original push for having a plug and play triplestore so people could try out different back-ends to see what works for them since performance issues
Seems to have been deprioritized with SDB performance improvements
Triplestore replacements will require extensive testing/evaluation
Performance upgrades in one context may yield differing results in other areas
Triplestore canonical data: the current architecture’s assumes this to be the case. Do we carry that forward?
If not triple-centric, then what does that mean? Practical implications? New code-base?
Linked data at input/output but spectrum of possibilities for internal architecture
Decoupling vs radical departure?
If swapping out triplestores, need testing. Technical challenge of writing the code is dwarfed by assessing whether this does what we want it to do?
If stepping away from triples, task got a lot bigger
Inclination to tease things apart - be able to integrate smoothly in ecosystem where triplestore is not necessarily canonical data
Inclined to hear about architectures where triplestores not “central”: what is the role of triples in this ecosystem
https://sul-dlss.github.io/rialto/architecture for reference
Inferencing is a blocking call
UI separation
Don is asked for general reflections:
For VIVO core group, incremental changes. Product evolution group: GraphQL, PostgreSQL etc…
Performance issues not terrible for the profiling system in his experience
Shortcomings of SDB to not handle "optionals" so have to do unions and constructs
Would like to see VIVO do linked open data. So close
With something like Blazegraph and/or performant triplestore, would want to do get all the data (not just public) - should be easy to export data and import it into another triplestore and query like a data warehouse
Since using Symplectic, not invested in features for bringing in external publications, etc. but can see that as useful and perhaps that needs to be a separate module
Mailing list messages
Freemarker template errors: that old nugget
Not all inferencing happening
Messages from agenda:
jquery.scrollTo error - Do we need to upgrade `jquery_plugins`?
No Subject - "How can I add more associated profiles for that particular editor."
CODE4LIB list: rdf and doi's - thanks Marijane White
Actions
Previous Actions
- Brian Lowe confirm LDF server issue with TDB content stores
- Don Elsborg - add jira tickets for abox/tbox use cases - one ticket for each use case
- Brian Lowe - check with ontology group on handles
- Alex Viggio will bring news of Elasticsearch instead of Solr up with Product Evolution.