The points that those who question the need for XSDs or RNGs are all
well taken, and they take into account the fact that schema is not just
a simple reworking of a DTD. Furthermore, I'm not *at all* suggesting
that one of the localized XSDs or RNGs be adopted for use by the
community at large. Because I've raised this issue before to officials
within the descriptive standards community, I've heard the same answers,
which essentially get into some sort of one-sided debate as to a) the
merits of DTDs, b) the fact that schema has only recently gained full
approval by the W3C, and c) how much time it would take to adequately
datatype EAD. These points are all well taken and are not falling on
deaf ears. But to some degree those that make these arguments are
overthinking this issue.
As Alejandro Delgado Gomez stated, I think most of us that have made
localized schemata would simply like to hear something official from EAD
officials on where the development of newer syntaxes stands. As I
understand, EAD officials have put some time and consideration into
altering EAD for datatyping, recoginizing the data typing issues. I
appreciate that effort, and would simply like to hear more about it. I
also wish that discussions about the percieved merits and problems with
all of the syntaxes could be amplified so that the rest of the community
could gain a better understanding of when we might (or might not) see
XSDs or RNGs. These issues certainly need to be discussed, as Stephen
points out.
The reason I want to hear where these discussions and efforts stand, and
also why I'm eager to see at least an EAD 2002 Schema, is that quite
simply the majority of the emerging standards and tools(!) that are
being utilizied within the digital library world are using schema as the
basis of their initiatives. It's true that many of the TEI-based
initiatives are still holdouts in adopting the newer syntaxes, but even
TEI-C now has a RNG. But the emergence of a schema for EAD need not
take years. Case in point: two equally voluminous and complex
initiatives -- the Federal Geographic Data Committee and the Data
Documnetation Initiative -- that were originally based on DTDs recently
adopted respective schemata to coexist with their DTDs so that their
participants could use datatyping and tools like JAXB, XForms, InfoPath,
etc., if they so desired. These schemata came into being after the user
community made their own localized schemas. The officials of the
respective initiatives took into consideration the users' needs and
either co-developed or adopted a user's schema in a time-efficient
manner. As stated, I'm in favor of the former over the latter.
Until the time comes that EAD or its predecessor can be reworked to
comply with schema, why not support a bare bones schema that can simply
let users in the community participate in larger inititatives that
require schema? I'm sure a survey of implementors will show that they
don't care if the date fields are datatyped for 8601 or Julian dates
(heck, it sounds like it'll need to be xs:string anyway). Other
communities are using other syntaxes for defining the data and are not
suffering because of it. Perhaps even more importantly, why not build a
schema so that people can build tools more readily to leverage EAD
metadata for emerging methods of storage (OORDBMS), authoring (re:
Terry's namespacing point for METS, MARCXML, MODS, and EAC references,
plus XInclude for inclusion of authority metadata), interchange via
harvesting (OAI), and searching (SRW/SRU)?
Keeping up with the Joneses and doing as the Romans do,
Clay
|