>>>>> On Sun, 23 Feb 97 21:12:59 EST, [log in to unmask] said:
Lee> Dave Beckett <[log in to unmask]> wrote:
Dave> [...] e.g the well known fake example: Element Date with value
Dave> 19970223 Qualifier name Scheme, qualifier value "ISO.1234(1996)"
Dave> gives:
Dave>
Dave> <META NAME="dc.Date.Scheme.ISO-2E1234-281996-29"
Dave> CONTENT="19970223">
Lee> Or, better, if somewhere (anywhere) you have:
Lee> <meta name="dc.scheme.iso.1234.1996"
Lee> url="http://www.iso.ch/standards/1234:1996">
Lee> you could then use:
Lee> <meta name="dc.date.iso.1234.1996" content="19970223">
No! The '.' is separating:
elements <-> qualifiers
qualifier names <-> qualifier AND
_part_ of a qualifier value
This is too much overloading of the same character. It also means
you can only have at most 1 qualifier since a simple program cannot
work out when a qualifier ends in your format.
e.g. for a Contributor's email address you need 2 qualifiers
Element: Contributor
Element value: [log in to unmask]
Qualifiers: Role=Editor, Type=Email
I see no one proposing that we ONLY need Scheme qualifers on each
element.
Lee> In practice, you'd probably not care about the date of the ISO
Lee> standard that much, and could write:
Lee> <meta name="dc.scheme.iso1234"
Lee> url="http://www.iso.ch/standards/1234:1996">
Lee> and then use <meta> name="dc.date.iso1234" content="19970223">
Lee> [snip]
Not only does this break all HTMLs with extra attribute URL and hence
we should not use it BUT again you have no way to specify 2 or more
qualifiers and the URLs for them.
Dave> Now the problem with this is that what happens when I don't want
Dave> to have a scheme, or scheme is unknown? In that case do I have to
Dave> leave a gap? ie. does dc.publisher.scheme.scheme-value.<MORE>
Dave> become: dc.publisher...<MORE>
Lee> No. If you don't have a scheme, leave it out: dc.publisher
The danger, as I mention above -- no way to have MORE qualifiers
Dave> In fact, having any requirement on qualifiers is going to fail.
Lee> I agree. It's necessary to allow unqualified fields for
Lee> unqualified authors :-) as well as tosimplify interoperability
Lee> where less precision is required.
Unqualified yes. Good (preferably with some semantics but we can't
guarantee this).
Skilled writers and DC-writing programs/converters will want to
include as much richness and detail that they can -- multiple
qualifiers, many elements, schemes, etc. and I can't see the method
above working for them beyond a simple Scheme qualifier mechanism.
Dave
|