Lists like MARC genre terms and relator codes, which are short and under
the LoC's control, can be included in the schema and automatically
updated if and when the LoC updates the list. Lists that are large or
open-ended can still be given a sanity check. Consider hypothetical
elements <isbn> and <uri>:
<xsd:element name="isbn">
<xsd:simpleType>
<xsd:restriction base="xsd:NMTOKEN">
<xsd:pattern value="[0-9]{9}[0-9X]"/>
</xsd:restriction>
</xsd:simpleType>
</xsd:element>
<xsd:element name="uri" type="xsd:anyURI"/>
Even this limited type-checking cannot be done when there are type
attributes as in <identifier type="isbn"> and <identifier type="uri">.
--Andy
>>> [log in to unmask] 02/03/05 1:42 PM >>>
Andrew E Switala wrote:
>The gist of it is that type attributes don't mean anything with
respect
>to validation. <genre authority="marcgt">abstract or summary</genre>
is
>valid, but so are <genre authority="marcgt">abstract</genre>, and
<genre
>authority="marcgt">marcgt my foot!</genre>.
>
This is true for any attributes, right? So what would our options be?
There are disadvantages to including the values in the schema itself,
especially if the values will be somewhat dynamic. Also, some of the
lists are very large.
kc
>
>
>
--
-----------------------------------
Karen Coyle / Digital Library Consultant
[log in to unmask] http://www.kcoyle.net
ph.: 510-540-7596
fx.: 510-848-3913
mo.: 510-435-8234
------------------------------------
|