ALEG

What I've done

• Two problems dominated work this week:
  1. After backing up and restoring the database, queries started running very slowly. After trying the usual steps to better inform the Oracle query optimizer on the contents of the data (and hence allow it to make a better query-path decision) with no success, I emailed an Oracle performance expert for help. His advice didn't seem to have any direct bearing on the problem (but it was very interesting background all the same!). But on trying his advice, the problem was fixed! Even more amazing, when I un-applied his suggested parameter change, the queries still ran just as quickly. So, I'm a bit confused, but I have some more resources to explore if (when!) the problem recurs.

  2. I encountered a problem using the Xalan XSL processor with a particular combination of circumstances trying to get the best performance out of Xalan. Fortunately, XALAN comes with source code (an open source package), so the problem was relatively easy to diagnose and fix. I've posted the fix to the Xalan developer group.

• Coded some more query enhancements, especially in the stylesheet area. The main purpose was not to produce a final user interface but to test Xalan XSL performance (which is good!) and to test the native XML production system of the ALEG Topic database (which will be required not just for the user interface generation, but also for data extract). This all went well.

• Kerry was here (Canberra) for ALIA, so spent time with her on Thursday/Friday discussing the issues with the BAL data load. There are many "issues"! Some she will address manually, others we'll try to automate whilst loading the data.

• Brief meeting with Anne Horn (with MLA, ALM, KK) updating project status.

• Exchange email with Judith and James proposing some Java interfaces to define the boundary with the Z39.50 holdings information retrieval subsystem.

• Document how ALEG will handle some tricky cases - The "Poets of the Month" works from the mid 1970's and "Down the Lake with Half a Chook". These are amongst the most "difficult" cases Tessa and Kathy can come up with, so if we think the proposed data model can handle these, we'll be happy!

• I'll be away Monday, Tuesday

• Start on the data maintenance (input) suite.

• Six months working on ALEG! Must be time for a hair-cut...