ALEG

What I've done

• Started the reimplementation of the results display user interafce. I actually started this late last week, but a power failure evaporated attempt #1...

  The reimplementation is most dramatic on the agent detailed display. Although the ancilliary attributes (bio, abstract...) are not yet being shown or linked to, counts of the works produced by the agent by type/form/genre are shown and hyperlinked. Also shown are works by type/form/genre which show the agent as subject (also hyperlinked) and a count and link to all the reviews about all the agent's works.

  The work summary and work detail screens have been revised, but not so extensively.

  I expect this will get a thorough review next week.

• Loaded the AUSTLIT agent/title as subjects. There are about 54,000 of these, and all but 2,000 have been loaded with some gentle coercion.

  These 2,000 are generally titles-as-subjects for which no work existed in AUSTLIT (and hence in ALEG). I guess these have to be manually checked out to create a work or discard.

• When reviews Works were created from the AUSTLIT review 'stubs' no "first known date" was available. I wrote a program to find the first known date based on containment within a publication and update the reviews.

• Kerry and Joan Keating noticed that very large works take a very long time to display in the Edit user interface. Lots of contained works are particularly problematic.

  After investigating the biggest cause of delay (which turned out to be the snazzy new Microsoft element behaviour components being used to make the prompt nicely hyperlinks) and rewriting some of the code, it now loads 33% faster. So a work containing 115 sub-works such as "Riding Out : New Writing from Around the World" now takes 28 seconds to load in the editor on my PC (Pentium Celeron 500Mhz). One of Peter Porter's collected works (with 2 manifestations and over 700 subworks and lots of reviews) takes nearly 3 minutes to load. This is a long time, but is a consequence of the "edit the whole work" approach we take. Further optimisation is possible, and I'll try to look at this soon.

  These large works are very much the exception. Typical single works and even average periodicals take only a few seconds to load.

• Sun Microsystems told us on Tuesday that they wouldn't be delivering the machine the previous day as promised, and there are now extended delays for some reason getting these machines.

  As we may not be getting the machine until March at the earliest, Fran is arranging with Sun to be loaned a machine on which we'll run the training and probably the initial production environment.

• Add more attributes to the work/expression/manifestation edit schema.
• Add general edits based on XSLT Schema (the Schematron approach).
• Add topic selection to the work/expression/manifestation UI
• Think about user access control

• Discuss and iterate the search and display user interface

• Keep improving the maintenance user interface, hopefully adding some custom templates to simplify entry of specific types of works (periodical issue, poem, etc)

• The hardware delay was a bit of a setback and distraction, but hopefully a loan machine will be arranged. Otherwise, things are coming together, but lots of work remains!