ALEG

What I've done

• The new PC's being leased by the ADFA library seemed as far away as ever, so Marie-Louise took me up on my offer to bring in my PC from home. It is only 4 months old, and has the speed (500Mhz Celeron) and memory (128MB) to run the Oracle client and a test Oracle database (although 128MB is a bit light-on to run an Oracle database of any size!).

  The PC had a Linux Red Hat 6.2 partition and a Windows 98 partition, but Oracle server requires NT, so I installed NT 4 Workstation, and then NT 4 service pack 6.

  It didn't have a Network Interface Card (NIC), and the only spare NICs available locally in the library were older ISA bus format cards, and being a cheap, new PC, it only supports the more recent PCI format card, so I bought a 10/100 ethernet card for $39 and installed it. Because I installed networking after installing the service pack, a few services didn't start (ah, Microsoft...), so after diagnosing the problem, I had to reinstall service pack 6. Interestingly, Linux automatically recognised and configured the NIC, but NT required a driver to be installed provided (provided with the NIC).

  Also, being a newer motherboard (Intel 810), NT 4 didnt recognise the sound card or video, so I had to find, download and install NT drivers for them.

  Finally! Installed IE 5.5 and Oracle 8.0.5. The ADFA computer centre does not have the media for Oracle 8.1.5 (the latest, which we want to use, especially because of the Oracle interMedia free text engine). So, we've ordered Oracle 8.1.5 for NT from the UNSW software distribution centre - we'll probably get it next week. I created a test Oracle database and table - all OK.

  Installed Java SDK 1.2.2. Unfortunately, Oracle 8.0.5 JDBC drivers are Java 1 and 1.1, not 1.2! So, I cant use the faster 'native' OCI drivers. So, until we install Oracle 8.1.5, I'll use the slow but simple JDBC to ODBC bridge for issuing requests to Oracle from Java. Tested this last thing Friday - simple requests worked fine, and it should be fine for loading data.

• The two 9GB drives for the Library's Sparc 450 have been ordered - when these arrive I'll transfer the Oracle database to these disks and stop running Oracle server on NT.

• Wrote some simple Java programs to analyse the BAL/LAW data in greater depth, which led to..

• Lots of emailed discussions with Kerry on the BAL and LAW data. I think we've pretty much discussed this to death, and the thing to do next is to load the AUSTLIT data and then check out how much overlap we have with BAL/LAW and see how our merging strategies will work.

• Discussions with Fran Cassidy about the AUSTLIT data to further refine the format she produces for loading into AUSTLIT - should get some data to load next week.

• Define the draft database and load some data to actually test our assumptions!

• 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!

• Instantiate the draft database design as Oracle tables

• Load the AUSTLIT data

• Load at least some parts of the BAL/LAW data

• Sanity-test simple, commonly used access paths to the data. Although the platform these tests will be run in is woefully short of memory, the CPU is fast (probably faster than the Sun 450), and it will give us an indication of the feasability of our proposed approach - if it won't work, it is better to know now than in 3 months!

• A week of much activity (futzing around with computer hardware and software) but little to show for it apart from lots more understanding of the BAL/LAW data thanks to Kerry.