Yes, having metadata in a repository certainly makes adjustments, global changes etc much easier. However it is essential for the metadata repository to have some mechanism for gathering information about changes to resources. It is not like a library catalogue or journal citation database where the resources, once published, do not change. >From the creator's point of view, it is easiest to record changes in the embedded metadata at the same time as the resource is changed. The metadata may in fact be routed to a number of gateway sites with metadata databases - you could not expect the creator to update the database in each of these repositories. On the other hand, it is easy to make errors when you are editing embedded metadata. This causes problems in gathering the updated metadata for the metadata repository (or repositories). There is no clear solution yet. Prue Prue Deacon HealthInsite Editorial Team Commonwealth Department of Health and Aged Care (Australia) [log in to unmask] phone: 61 2 6289 7505 "Dwight Walker" <[log in to unmask]> on 29/10/99 18:29:29 To: [log in to unmask], [log in to unmask] cc: [log in to unmask], [log in to unmask] (bcc: Prue Deacon/NHMRC/Health) Subject: Re: From embedded metadata to repository I was at the NSW Parliamentary Library today at a talk given by Kath Curr, the librarian, for ALIA ISS. They are implementing a state-wide intranet to publish clippings. I asked if they used meta data. They said yes but not Dublin Core as the system is totally closed off from the general public so does not need to conform to external standards. The choice of meta data headings is necessary. You may customise this based on your circumstances. Next I know Oracle's new database will extract meta data from multimedia like CDs directly. You may be able to automate the extraction of meta data depending on your media - text, PDFs, Word, Excel, audio, video. Once it is in a proper database or repository you will be laughing. It is much harder to hand-edit the HTML than use a database and some language like ASP or Perl to generate HTML with the meta data codes embedded in it. I'd go with a repository and use some text manipulation language like Perl to extract meta data from existing Web pages to insert it in there. I was reading up some Perl last night. It can suck in the meta data for you into an SQL database. It is a very powerful text processing language. Why not get a Web developer to do this for you? Web programming is much less technical than traditional software development. > > For example, I am have been told that it would not be difficult to extract the > > embedded metadata at a later stage once a repository is available. Any comments? ---------------------------------------------------- Dwight Walker WWWalker Web Development Pty. Ltd. (ACN 088 959 086) PO Box 428, Merrylands, Sydney, 2160, Australia http://www.wwwalker.com.au, http://www.speakeasy.org/~dwight (US mirror) tel +61-2-96371649, mob +61-412-405727, fax +61-2-97772058 ICQ No. 4631678, handle: wwwalker (www.icq.com) %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%