VuFind Developers Call Minutes: August 20, 2013

Attending: Filipe Bento, Chris Hallberg, Demian Katz, David Maus, Tod Olson, Sean Purcell, Dan Scott, Benjamin Wiens

Note: JIRA has now been updated to integrate with GitHub, so you can see commits related to tickets and in the status feed.

• VUFIND-815 - SolrMarc has been upgraded to the latest version.
• VUFIND-856 - Demian has committed replicated index support to the Linux version of the alphabrowse generation script; Windows support is still needed but will not be worked on unless requested.
• VUFIND-868 - Summon cover image loading has been updated to reflect a minor API change.
• VUFIND-869 - This Horizon driver bug fix has been committed.
• VUFIND-870 - This ticket proposes an “asset pipeline” for streamlining delivery of CSS and Javascript resources to the client.
• VUFIND-871 - This ticket proposes cosmetic improvements to the advanced search screen. Chris' bootstrap theme will help with the problem.
• VUFIND-872 - This ticket proposes an extra PHPUnit parameter to prevent “out of memory” errors. Demian will commit if he can get it working (no luck so far).
• VUFIND-873 - This ticket covers a bug fix (now committed) to correct bad URLs generated by the driver-based “new items” and “course reserves” searches (they were including unnecessary, and sometimes very long, lookfor parameters in the facet links).
• VUFIND-874 - This ticket reports a notice-level error in the statistics system; Demian has fixed it.

Now that we are on GitHub, we should start going through important pull requests; Demian will omit very small issues like translation tweaks and configuration comment adjustments.

• #7 - Breadcrumbs in the author module have been made more complete.
• #9 - The publication date graph recommendation module is now hidden by default (rather than always displaying and then hiding itself when data is missing); this makes the user experience slightly better, especially if JS is disabled.
• #11 - This patch allows more than 30 facet values to be loaded dynamically into the facet list; it needs a bit more polishing before being pushed into master.
• #13 - This is a CAS authentication implementation (see also VUFIND-422).
• #16 / #17 - These PR's fix catalog password encryption issues for the Shibboleth and LDAP authentication modules; see also VUFIND-866 / VUFIND-867. Demian will test and merge soon.
• #19 - This is a work in progress to upgrade the jquerymobile library (see also VUFIND-844).
• #20 - This is based on the QR code support in VUFIND-308; there is still some ongoing conversation about details.
• #21 - This contains a schema.org RDFa implementation related to VUFIND-425.
• #22 - This request (now merged) contains some minor optimizations/adjustments to the search system for the 2.1 release.

Chris is still working on this – hope to have something to share soon after 2.1 release.

Discussion question: following release 2.1, should we create a 2.1 branch for bug fixes and merge them back into master, or should we create a 2.2 branch for major new development and merge master fixes into that branch?

General consensus: keep bleeding edge development on master, create release branches to backport important bug fixes.

Dan asked about release schedules (time-based, or feature-based?). Demian explained that right now VuFind scheduling is largely feature-based, but time-based is not off the table for the future if that model becomes more appropriate.

Dan elaborated on his notes from the JIRA ticket and PR related to RDFa. He highlighted the movement to use the schema.org “Offer” (originally designed for commercial purposes) to expose holdings data.

Tod asked whether the patch is “on by default” – Dan said that yes, it is, but its footprint is fairly light.

There was some discussion about where holdings metadata should be exposed (just on the holdings tab? everywhere? if everywhere, what are performance implications?). Initial approach is to only expose on holdings tab and hope that is sufficient.

There was discussion of whether record driver subclassing might cause problems in relation to the RDFa – it seems unlikely in the current model (or at least not to be more difficult than other types of changes).

Also a question of when this can be accepted – currently hoping to hold off until after Bootstrap theme is introduced to simplify keeping things in sync; Dan offered to rebase onto Bootstrap theme once it is ready.

Dan will add some notes to the JIRA ticket regarding tools for testing RDFa.

Dan also mentioned that he may be able to work on testing/improving the Evergreen driver.

David mentioned that we should begin planning our goals for the 2.2 release; this should be the main focus of the next call.

The call for papers has gone out; registration will open shortly.

Filipe set up an instance of Big Blue Button, and we experimented logging into it. Some users had problems with Flash plugins crashing, and desktop sharing took some time to set up (but eventually worked). Flash issues may be specific to the Chrome built-in plugin.

We'll have to discuss further on the mailing list to decide if this is a workable solution.

The next call will be Tuesday, September 3, 2013 at 10am Eastern Daylight Time (14:00 GMT).