ALEG

What I've done

• Happy to be back working on ALEG after a month off. There were 3 separate problems of data inconsitencies during April caused by system bugs which were the main priority for this week. I think I've located the cause for these and fixed them.

  The most insidious problem was caused by an expression or manifestation entry in a data maintainers' "most recently referenced" lists remaining available for selection even after the work it was part of had benn deleted from the database. Or even more subtlely, an expression or manifestation could be chosen as a source/contained-within relationship for a work, but before that work was saved, in a second window that chosen expression or manifestation could be deleted.

  The server did not check that the expression or manifestation was valid, and hence the database was inconsistent.

  Although I've fixed this problem, other instances could occur (eg, select an agent in one window, delete the agent in another window, save the first window), and I must systematically check all topic references for validity at the time a work/agent is saved, not just at the time of selection.

  Other problems were the occassional duplicatation of an event (birth, embodiment, etc) and the occassional deletion of a date topic which was in use (and shouldn't have been deleted!).

• Many minor changes to formatting, templates, edit checks

• Merge works. This task will build on the infrastructure used for nerging agents, but needs to add support for merging expressions and manifestations, and moving manifestations from one expression to another

• To Do list. Work through the (mostly) minor issues on the to-do list provided by Annette/Kerry/Marie-Louise. I'll try to report on what I've done from this "background" list each week, and I expect it will be revised and reprioritised often!

• Brief overview presentation of technical aspects of the ALEG project to the NLA.

• Lu Rees data load. I was working on this in March, and the load program is very close to working - needs testing.

• First known dates. Our strategy of calculating first-known-date (fkd) at the work level has proven inadequate. For a simple example of its failure, consider an anthology which is revised into another expression and includes a new work: currently the fkd of that new work is calculated based on the fkd of the anthology, not the new expression, so the new work *may* have a fkd of before the birth date of the author! I think we have to record fkd of expressions as well as works.

• The new thesaurus is expected to be available for loading soon.