Community Call today at Noon ET

[Danny Brooke]

Hi everyone,

We'll have our regular call at Noon today. Connection info and notes document:

https://dataverse.org/community-calls

See everyone soon!

- Danny

[Philip Durbin]

Great call today. Here are the notes from https://docs.google.com/document/d/1pKZvMBfNeURPmKM0rMLSDm1a-0oF5I-PfPGemrM9dEc/edit?usp=sharing

2019-09-24 Dataverse Community Call

Agenda

* Release Updates, Release Notes
* Community Questions

Attendees

* Danny Brooke (IQSS)
* Gustavo Durand (IQSS)
* Tania Schlatter (IQSS)
* Sherry Lake (UVA)
* Jim Myers (QDR, TDL)
* Laura Waugh (TXST, TDR)
* Jamie Jamison (UCLA)
* Phil Durbin (IQSS)

Notes

* (Danny) 4.17
   * Next week or the week after.
   * Dataset level explore tools
   * Changes to user session timeout to help with performance
   * Around 30 other smaller items
   * New release notes process, under version control
      * If you want to help, jump on the branch linked from https://github.com/IQSS/dataverse/issues/6185
   * Any questions
      * (Jim) The session stuff sounds like a mysterious increase in memory use. Have talked about setting up cron jobs to restart Glassfish to release memory.
         * (Danny) We do expect a big change.
         * (Gustavo) Better performance with the memory going down. We not sure if we should call it a memory leak. We're trying to get away from using "view scope" when it's not needed, using "request scope" instead.
         * (Phil) Anonymous sessions have been reduced from 24 hours to 10 minutes.
         * (Danny) More details coming in the release notes.
* Community Questions
   * (Jim) Questions about Make Data Count
      * Who’s using it now?
         * (Danny) No one that I know of. Harvard Dataverse is using this issue to track setting up Make Data Count: https://github.com/IQSS/dataverse.harvard.edu/issues/3
         * (Phil) There's a group at Purdue that opened a support request about Make Data Count and Counter Processor: https://help.hmdc.harvard.edu/Ticket/Display.html?id=280620
         * (Danny) We want to make sure the counts aren't zero when we switch over the metrics display.
         * (Sherry) Is it documented?
            * (Phil) Yes, but we found a few issues: https://github.com/IQSS/dataverse/issues/6082
      * Is there MDC work going on now? E.g. on more detailed displays, including pre-MDC data counts in some way?
         * (Phil) In the future, I think we should dream big about
      * Would it be useful to allow metrics to be collected without switching the display (e.g. let instances start collecting now and switch later when they have the time to set up Counter/when more detailed displays arrive, etc.)?
         * (Jim) Maybe we should have different database settings for telling Dataverse to start logging vs displaying the Make Data Count metrics in the UI. I'll make an issue.
            * (Danny) Thanks!
            * (Jim, later) Done: https://github.com/IQSS/dataverse/issues/6212
   * (Danny) New SDK, new API client library for Dataverse for Javascript: https://groups.google.com/d/msg/dataverse-community/vnbVAmgcnvM/1cfS-bAfBAAJ . A group here at Harvard is developing it for a project having to do with optical character recognition and machine learning.

--
You received this message because you are subscribed to the Google Groups "Dataverse Users Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-commu...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/904223d7-4872-4260-a658-b9e3ef6b9034%40googlegroups.com.

--

Philip Durbin
Software Developer for http://dataverse.org
http://www.iq.harvard.edu/people/philip-durbin

[Janet McDougall - Australian Data Archive]

hi All

ADA has a project  to recover historic population and demographic data (tabular) from printed volumes, so we have been doing some OCR work using Tesseract with initial test results looking promising, but we have only just started.. 

What sort of work is the group you mentioned doing? 

"A group here at Harvard is developing it for a project having to do with optical character recognition and machine learning."

thanks

Janet

[Philip Durbin]

Hi Janet, I don't know a lot about the OCR project but from a technical perspective I think it's interesting that in https://github.com/dell-research-harvard/Break-the-Page they are detecting text blocks in a scanned page for further analysis as in the attached screenshot.

--

You received this message because you are subscribed to the Google Groups "Dataverse Users Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-commu...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/2d02efbc-3647-46ee-b356-73c562a47e81%40googlegroups.com.