OLAC Schema version 0.4 released

Steven Bird sb at UNAGI.CIS.UPENN.EDU
Tue Oct 23 10:28:20 UTC 2001


Dear OLAC implementers,

The new OLAC Metadata Set document has just been announced on the
OLAC-General mailing list.  As Gary said, we're also releasing the
corresponding XML schema (version 0.4):

    http://www.language-archives.org/OLAC/0.4/

Please note the following files in particular.

The top level schema:
    http://www.language-archives.org/OLAC/0.4/olac.xsd

An example record:
    http://www.language-archives.org/OLAC/0.4/olac.xml

Some of you reported problems with 0.3 with respect to capitalization
(the schema was not consistent with the official definitions).  The
capitalization of element names and attribute values is now consistent,
and follows Dublin Core.

If you produce OLAC records conforming to 0.3, the minimal changes you will
need to make to get 0.4 conformance are as follows:

(i)   make all element names lowercase;

(ii)  lowercase the first letter of the following attribute values:
      * refine attribute of the creator element (OLAC-Role)
      * refine attribute of the contributor element (OLAC-Role)
      * refine attribute of the date element (OLAC-Date)
      * refine attribute of the relation element (OLAC-Relation)

(iii) change the type.data element name to type.linguistic

(iv)  change the lang attribute of the olac element to langs.


I append a full list of changes since version 0.3.

      1. Capitalization rules follow Dublin Core
         (a) metadata element names are no longer capitalized
         (b) vocabularies OLAC-Role, OLAC-Date, OLAC-Relation
             are no longer capitalized
         NB. other vocabularies with initial capitals are unchanged:
         * DC-Type (following Dublin Core)
         * OLAC-OS-Code (use of capitals seems right,
           e.g. Unix/SunOS, OS2)
         * OLAC-Sourcecode-Code (language names are conventionally
           capitalized, e.g. Lisp)

      2. A new optional "scheme" attribute has been added on all elements,
         to support external control of element content (e.g. for controlled
         vocabularies over dialect names; see http://arXiv.org/abs/cs/0110014)

      3. The OLAC-Data controlled vocabulary has been renamed to
         OLAC-Linguistic-Type, in response to the UCSB workshop.
         The corresponding metadata element is called type.linguistic
         (was type.data).

      4. The included schemas now have absolute pathnames (this fixes a
         problem people reported for xerces).

      5. The lang attribute of the olac element has been changed to langs,
         since the element content of a record could be in multiple languages.

      6. The title element now has an optional refine attribute which may
         take a single value "alternative" (following DCQ).

For full details, examples and discussion, please refer to the new
OLAC Metadata Set document at:

    http://www.language-archives.org/OLAC/olacms-20011022.html

Please contact me if you have any questions.  As always, we welcome
your feedback and suggestions.

Thanks,
Steven.

________
Steven Bird, University of Pennsylvania (sb at ldc.upenn.edu)
Gary Simons, SIL International (gary_simons at sil.org)
OLAC Coordinators (www.language-archives.org)



More information about the Olac-implementers mailing list