ALEG

What I've done

• Two frustrating days spent getting the Oracle 8i Intermedia free text searching working under Solaris. Attempting to creating text indices using Intermedia failed with an error message indicating that the Oracle "extproc" (external procedure, to which text indexing is delegated) could not be launched. This problem was common reported on internet mailing/discussion lists, so I worked through the various solutions offered, but none fixed the problem. Eventually the problem was tracked down to a domain suffix being added by the service resolver which attempted to find the extproc, but not being defined in the service configuration file created by the Oracle Installer - it turned out to be a very simple problem caused by the Installer, but I was looking for complicated problems caused by running 2 versions of Oracle on the same machine...

• Started to tackle the problem of matching AUSTLIT sources with AUSTLIT titles. AUSTLIT stores sources and titles quite separately, but in ALEG we want to match the source to a work. Unfortunately, this is difficult because the representation of the source title rarely exactly matches the representation of the AUSTLIT work title - spelling variations plus title variations (use of alternative title), author variations (use of alternative names), date variations conspire to make automated matching either very difficult or risky.

  There are 9592 AUSTLIT sources. Of these, 5580 have no author specified (mostly periodicals not recorded as works in AUSTLIT). Of the 4012 with authors, 3557 have now been matched with AUSTLIT derived expressions, leaving 455 to be matched. Some of these are non-AUSTLIT works, some are just gross mismatches which will need manual assignment.

  I'm currently working on the 5880 without authors, trying to recognize the genuine periodicals and create ALEG works for them. Some are not periodicals (instead, they are monographs or collected works without authors specified), many of the 5880 are spelling variations which I'm trying to unify.

• 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!

• Finish matching the sources (all that can be done automatically) and creating all the part manifestations by processing the AUSTLIT sources file to create/match compound works and expressions, deal with the periodical sources such as The Bulletin, The Age etc

  Fran will also be producing a file of the reviews, so we'll have to see how best to process them.

• More frustration with the Oracle installation and lots of data matching problems with sources, but I'm whittling away at them, and getting to know the data much better than I thought I would!