AustLit

What I've done

• Very little progress on implementing the scoped search or converting the works with awards-as-subjects and dates to use scoped awards as subjects. Some searching algorithms have been designed and tested, and they seem to work, but I haven't yet added them to the search subsystem.

• A great deal of time has been spent in email and meetings discussing issues arising from our implementation of the FRBR model, especially:
  • use of the first-known-date and the override date, and the use of dates on creation, realisation and embodiment events

  • when the changes to an expression are so great that they trigger definition of a new (but related) work

  • pseuds/writing names, and especially whether attributes transfer from the "real" person to the pseud (and vica-versa), and does a search for works by the pseud return works by the "real" person (and vica versa).

  I think some of the discussions are arising from the consideration of the scope constraints of AustLit, which doesn't seek to be a universal catalogue of everything, but has a strong focus on Australian Literature. Consequently, music, songs, illustrations, performances and other non-literary expressions, although perfectly representable within our FRBR model, are not "core", and when non-core attributes (such as dates) effect the literary side of things, well, it gets ugly. For example, if a poem existed/was published as a "song" (say, in sheet music) before it was published as a poem, should that date be taken as the first-known-date of the poem? If a poem is slightly illustrated, then it is probably just a new expression of the poem, but at some point, the illustrations are so extensive that is morphs into a new work, maybe a "picture book".

  I suppose creating a specialised set or view of the world will always bring up these problems, especially if the broader goal is for interoperability with broader systems (such as a national/international catalogues which may one day be based on FRBR): it would probably not be good if we treated something as an expression, yet general cataloguers treated the same thing as a separate work...

• Produced two papers for the Ausweb02 conference in July:
  • Schema Driven User Interface Generation - Building Advanced Web-based Interfaces using presentation schema, XML, the DOM and background HTTP connections

  • Taking RDF and Topic Maps seriously - what happens when you drink the Kool Aid.

  AustLit are paying for my registration and I'm paying for travel and accomodation.

• Marie-Louise has prepared a fantastic paper for the IFLA conference later in the year: Report on the successful AustLit: Australian Literature Gateway implementation of the FRBR and INDECS event models, and implications for other FRBR implementations.

• Changed all role=author/creator to role=illustrator on works of form "interview" where there was just one author or creator.

• Changed all role=sponser to role=publisher on works of type "publisher series".

• Added new form type: "index".

• Added 'publication year' as a new search criteria on advanced search following a request from Kerry, but I'm not sure this is what is wanted as it only searches on manifestation year, and doesn't include source year...

• Changed the "First Known Date" calculation to ignore work creation date

• A few changes to programs, stylesheets and other formatting following maintainer and user feedback (suppressed language at the expression level of works being shown as a source); sundry template changes (maint note now defaults on new agent/work, publisher removed from periodical); changed author-creation-date rendering as suggested by Joan; bundled up manual notetab files and DTD's for Carol; )

• The system was stable.

  However, this morning we discovered that ADFA ITSC had implemented a new firewall arrangement, and that the public IP address of the AustLit server had been changed!! Although the DNS name remains the same (www.austlit.edu.au), DNS changes take time to propogate through the network, and hence many users could not connect to AustLit as the server name was been mapped into the old IP address. To compound matters, there was a problem with the firewall translating the public IP address of the AustLit server into the "private" IP address, which prevented access from users even with the new DNS mapping, but this was quickly fixed by ITSC. So, a frustrating morning, and apologies to our users for the unavailability of AustLit for several hours.

  I've had many reports from UQ and Flinders over the last fortnight of problems accessing AustLit, or very slow response. However, the server has been lightly loaded, and I cannot reproduce them. Often, one remote site is having difficulties, and others are working fine, so I suspect that most of these issues are related to wide-area-network problems, which we also experience here quite frequently trying to get to outside sites.

Next Fortnight

• Conversion of award-as-subject with years to use relationship scopes. Concurrently change search algorithms to use the scope for year searches associated with awards.

• Think about preparing AusWeb02 presentations.

• Move awards to be event-based rather than work-based.

• Refining NBD Holdings searches.

• Review all subset definitions for efficacy.

• Import/export in MARC and DC formats.

• SDI facility.

• Combining searches

• System Documentation