<HTML>
<HEAD>
<META content="text/html; charset=big5" http-equiv=Content-Type>
<META content="Open WebMail 2.32 20040525" name=GENERATOR>
</HEAD>
<BODY bgColor=#ffffff>
<font size="2">
<p><font size="2">Dear All:</font></p>
<p><font size="2">I agree witih Heidi.</font></p>
<p><font size="2">Chu-Ren</font></p>
On Thu, 30 Sep 2004 16:44:13 -0400, Heidi Johnson wrote
<br />> Nick Thieberger wrote:
<br />> >Dear OLAC implemeenters
<br />>
<br />> >I write to ask if the metadata discusssion is ongoing or is it now
<br />closed? As we
<br />> >develop our archive we are adding to our controlled vocabularies and
<br />wonder if the
<br />> >OLAC (that is, the community) want to add these to the standards, or
<br />prefer to have
<br />> >metadata extensions for each small repository.
<br />>
<br />> >It seems to me that enhancing the OLAC standard would be preferable, so
<br />we are
<br />> >suggesting that 'song' and 'instrumental_music' be added to the
<br />datatypes.
<br />>
<br />> >We also ask that 'isTranscriptof', 'isTranscribedby' be added as
<br />attributes of the
<br />> >relation element.
<br />>
<br />> What did we decide about a revision schedule for metadata terms? Did
<br />> we decide anything? Maybe we should have something like an annual
<br />> revision protocol, whereby we discuss revisions on the list for a
<br />> month or so and then adjust the docs? Would that make sense?
<br />>
<br />> We certainly need to change the controlled vocabularies from time to
<br />> time. I remember that our goal during the workshops was to keep the
<br />> lists as short as possible, but my personal feeling is that they
<br />> should be allowed to grow as needed. 100 terms is a lot to display
<br />> on a menu, but it's not particularly troublesome to search. Archives
<br />> might not need all of them, but as a group we will surely end up
<br />> needed a largish set of, for example, genre terms.
<br />>
<br />> I've already added 'chant' to AILLA's genre keywords, even though
<br />> it's not in either OLAC or IMDI. But every second thing from
<br />> Amazonia is a chant, and all the experts agree that a chant is a
<br />> chant is a chant: it's not speech+song, or whatever. We also have
<br />> 'song', so I second Nick's amendment, and agree that we need
<br />> 'instrumental_music' as well.
<br />>
<br />> I think we were being too strict about whether a genre is linguistic
<br />> or not. In documenting endangered languages, we get all kinds of
<br />> materials that properly belong in the same corpus, although they may
<br />> not be linguistic data. instrumental music is certainly one such;
<br />> teaching materials are another. We are getting more and more of this
<br />> kind of thing (calendars, illustrated encyclopedias, primers, etc.)
<br />> at AILLA and we strongly encourage archiving and disseminating
<br />> these things broadly, because they are so useful for helping
<br />> everyone think of better ways to revitalize and support these
<br />> languages. So, we need genre terms for this stuff, too. Any ideas?
<br />>
<br />> I haven't seen any attributes of the Relation element: is there a
<br />> document for that? I would also second Nick's 'isTranscriptof',
<br />> 'isTranscribedby'. Perhaps then we would also want
<br />> 'isTranslationof', 'isAnnotationof'?? The standard companion texts,
<br />> in other words.
<br />>
<br />> I think it's worth being fully descriptive rather than strictly
<br />> parsimonious with terms.
<br />>
<br />> Just my $0.02.
<br />>
<br />> Heidi Johnson
<br />
<br />
<br />=======================================================
<br />Chu-Ren Huang (churen@sinica.edu.tw)
<br /><a href="http://corpus.ling.sinica.edu.tw/member/churen/" target="_blank">http://corpus.ling.sinica.edu.tw/member/churen/</a>
<br />Institute of Linguistics, Academia Sinica
<br />Nankang, Taipei, Taiwan 115
<br />ph. 886-2-26523108
<br />fax. 886-2-27856622
<br /></font>
</BODY>
</HTML>