Architecture for Access to Government Information (Australia)
Fri, 27 Sep 1996 10:33:37 +1000
robert @ lyvvie.virago.org.au wrote 26/09/96 08:53 PM:
>On Thu, 26 Sep 1996 Tom_Worthington@dmd.a-l.defence.gov.au wrote:
>> As one of the members who wrote the report I will try and help:
>I am interested in what indexing schemes the committee considered
>and why they chose the Dublin Core in preference to some of the
>others that are doing the rounds (such as GILS). Any comment?
Metadata options are considered in chapter six of the report <
http://www.adfa.oz.au/DOD/imsc/imsctg/imsctg1a.htm> and were:
=F8 GILS Profile
=F8 Dublin Core
=F8 TEI (Text Encoding Initiative) SGML DTD
=F8 IAFA templates
It really came down to GILS v Dublin Core and in fact the report
recommends a subset of GILS to be called AusGILS (pronounced OZzzGills).
The priority for the group was to consider metadata for collections of
documents, rather than individual documents. As GILS was designed for
that purpose it was considered better for collections.
For individual government documents the report is a bit vague, suggesti=
AusGILS, the Dublin Core, or an appropriate specialised metadata set.
However Dublin Core appears to have been gaining ground since the repor=
I am an computer person and the selection of which set of meta-data ite=
to use was of less interest to me, so I left that to the librarians,
records managers and archivists. My interest was to see that we had
something which could be easily implemented quickly using available
Any of the metadata sets could be encoded, stored and searched in a
number of ways. The minimalist approach I proposed in Appendix D was th=
use of the HTML META tag in Web documents. When first proposed this was=
radical idea, compared to using a GILS database approach. However we so=
discovered that it had already been proposed and was being supported
The refinement of this approach which I proposed to the searching
workshop at NLA <http://www.adfa.oz.au/DOD/imsc/imsctg/workshop/> was
that the "Web Crawler - Web Search" and "database" options be merged.
That is meta data could be encoded in individual Web documents and woul=
then be collected by Web crawlers or it would be manually entered into
specialised databases. In either case it would be available to be
searched as if it was harvested from Web documents via a web type
interface or as if it had been entered in a database via a database typ=
The concept of a "document" becomes a very abstract in this environment.
Some agencies are using databases to serve up Web documents. Each
document is generated on the fly from components in a database (this
could include metadata) and sent as HTML. The receiver may be a program
which breaks the document back into its component parts for analysis. N=
one actually sees a document which looks anything like a piece of paper.
Tom Worthington <Tom_Worthington@dmd.a-l.defence.gov.au>
Deputy Director Information Management Plans
& Defence Web Administrator
Information Management Branch
Department of Defence <URL: http://www.adfa.oz.au/DOD/dodhmpgn.html>
Room NBH-6-06, Canberra ACT 2600, Australia
Ph: +61 6 2665450 Fax: +61 6 2665311 AH: 0419 496150