The differences are largely quite technical, for discussion of this I'd 
recommend Rick Jelliffe's still helpful "The Current State of the Art of 
Schema Languages for XML" ("current" being 2001). Some of the most 
significant differences relevant to EAD application are:

* datatyping

* internal annotation

* namespaces

W3C XML schema and RelaxNG have stronger features for the specification of 
datatypes than does DTD. So, for example, a date can be identified as a 
date datatype and processed as a *date* -- for example with XSLT2 
functions like format-date() which allows the specification of the format 
a date should be output as (YYYY-MM-DD; Year, Month Day; Month Day, 

Along these lines, with W3C XML and RelaxNG, values for attributes and 
elements can be more tightly defined. This is when validated against the 
schema versions of EAD, the ISO country codes in a DTD-valid instance may 
cause errors. In the schemas the codes have been defined strictly as to 
case, while this is not the case in the DTD. Similar constraints might be 
very desirable if one wants to create a subset/local-use version of EAD.

W3C XML Schema and RelaxNG also allow the embedding of textual annotation 
in the schema itself. This is a useful feature exploited by software like 
oxygen which will display the documentation for an element from the 
annotation at the users request. Again, this feature can be useful in 
subsetting/local-use schemas where additional usage guidelines and 
element/attribute definitions can be included within the schema.

I believe it was the desired for namespace functionality which provided 
the initial motivation for a schema version. In both W3C schema and 
RelaxNG it is permissible to include elements from other schemas in your 
documents. The most relevant example is METS. A METS file uses data in 
external schemas to provide metadata digital objects it 
describes/instantiates. One of the recommended schemas is EAD. However, 
because there had been no official schema version of EAD, EAD data could 
not be easily (or "officially") validated inside METS. Now, it can.

Hope this helps. Perhaps others can add to this or clarify as well.


On Thu, 1 Nov 2007, Russell D. James wrote:

> So answers to my initial question bring up another query from me. Just what
> is the difference in operation between the two schema and which one is used
> more often than the other?  Is there a preference?  Why two different
> schemas?
> Russell D. James, CA
> [log in to unmask]

Terry Catapano
Special Collections Analyst/Librarian
Columbia University Libraries Digital Program
[log in to unmask]