This technology could take one of two biases:
1. Structured values are typically represented using a private syntax
specific to the type of data.
E.g.
<DC:DateRange>
1997-01-01/1997-12-31
</DC:Date?Range>
2. Structured values are typically represented using a standard syntax
such as XML. If so, a standard parser can automatically discover the
structure.
E.g.
<DC:DateRange>
<DC:DateMin>1997-01-01</DC:DateMin>
<DC:DateMax>1997-12-31</DC:DateMax>
</DC:DateRange>
Date ranges are only one example. Others are lists, collections, people's
names, addresses, phone numbers and so on ad infinitum.
--Andrew Layman
[log in to unmask]
> -----Original Message-----
> From: Eric Miller [SMTP:[log in to unmask]]
> Sent: Thursday, October 30, 1997 8:23 PM
> To: Andrew Layman
> Cc: 'Dirk-Willem van Gulik'; Misha Wolf; meta2; RDF Syntax (E-mail)
> Subject: RE: Date ranges
>
> Andrew Layman writes:
> > Does Dublin Core prohibit explicitly structured values?
>
> No, Dublin Core does not prohibit structured values.
>
> The Dublin Core focus has generally been on semantics. Tools and
> implementation utilizing the benifits of shared semantics, however,
> invariably grapple with issues of how to tranfer this stuff
> around. One of these syntaxes is HTML. HTML (specifically the
> content-less META) is a syntax that does not lend itself well to
> structured description...
>
> I take Dirk's point a suggestion of how best to represent structured
> Dublin Core, in essence, within the limitation of HTML META.
>
> Please continue this thread, however, on the meta2 list, not on
> w3c-rdf-syntax-wg.
>
> thanks,
>
> ---
> eric j. miller OCLC Online Computer Library Center
> mailto:[log in to unmask] Office of Research
> http://purl.oclc.org/net/eric phone:(614)764-6109
>
> >
> > --Andrew Layman
> > [log in to unmask]
> >
> > > -----Original Message-----
> > > From: Dirk-Willem van Gulik [SMTP:[log in to unmask]]
> > > Sent: Thursday, October 30, 1997 1:22 AM
> > > To: Misha Wolf
> > > Cc: meta2; Andrew Layman
> > > Subject: Re: Date ranges
> > >
> > > On Wed, 29 Oct 1997, Misha Wolf wrote:
> > >
> > > > focus on metadata encoded using HTML's meta elements. Please
> > > remember that
> > > > in XML (and, hence, in RDF) it will be easy to write something like
> > > (it
> > > > probably won't look precisely like this):
> > > >
> > > > <DC:Date>
> > > > <DC:DateMin>1997-01-01</DC:DateMin>
> > > > <DC:DateMax>1997-12-31</DC:DateMax>
> > > > </DC:Date>
> > > >
> > > > Comments?
> > >
> > > Well, what I see above is ONE information element; which conveys a
> > > date range to me. And I absolutely agree that this is a nice syntax
> > > as the XML allows for such tight groupings. However the currently
> > > avaible language does not allow for such tight grouping of two
> > > or more 'components' which together make up one dublin core element.
> > >
> > > So I would still strongly advocate to only convey one bit of
> > > information
> > > per element; not a half, and not two; so IMHO
> > >
> > > NOT Keyword: water, artic, algea (three elements)
> > > NOT Date.1.min: 1997-01-01 (half element)
> > > NOT Date.1.max: 1997-01-04 (half element)
> > >
> > > but Keyword: water
> > > Keyword: artic
> > > Keyword: algea
> > > Date: 1997-01-01 1997-01-05
> > >
> > > Adhering to the simple dogma that one element conveys one piece of
> > > information, then little syntaxtic sugar is needed; the above
> > > can !only! mean a range; where it two points in time one would
> > > have had to write
> > >
> > > Date: 1997-01-01
> > > Date: 1997-01-05
> > >
> > > So what is left to solve is an open ended range.
> > >
> > > Date: 1997-01-01 *
> > >
> > > Where '*' is any token which does noet look like a valid dat :-)
> > > I'd use something like a `+` or a `>` , '...' etc..
> > >
> > > The inexact dates are equally complex for both the 1.max way of
> > > writing as for this one-on-one kind of writing.
> > >
> > > Dw.
> > >
|