standards

Iris Jastram wrote:

> Just my two cents, but I think that one major push we'll have to
> go through in the coming years will be to come up with the
> standards on which all the possible component parts of an MLS
> (modular library system) can be built.

I concur. Standards are the way to go.

Libraries are not so special that they need to have this gargantuan
thing called the ILS, per se. Just like desktop software, you can
purchase the Microsoft Office Suite and and get all of this (extra)
stuff, or you can get a text editor, a spreadsheet, a database, and
image editor, etc. to create a different whole.

The "next generation" needs to adhere to standard, best-practices
input and output. Let's begin with MARC. Okay, it is embodied in a
weird data structure designed for tape, but the content is really
useful. Let's first use that. Then there is all sorts of content
available in other metadata formats such as OAI. Harvest an OAI
server containing the content you think is important to your patrons
and include that in your user interface. Then there are full text
things like TEI files or EAD files. Pull those standard things in
too. Mirror open access journal content, convert it to XHTML, and add
that to your "catalog".

Store the data in a relational database using standard SQL. No MySQL-
isms. No Oracle-isms. This will enable you to move from one database
to other database with ease. Index your content. There are no
"standard" indexing engines, but you can standardized on what is
indexed and how it is searched. For example, you can put a Z39.50
interface, or preferably in this Webbed world an SRW/U interface on
the index. This will provide a standard way of getting content in and
out of the index. SRW/U and Z30.50 will give you control of search
lists.

OpenURL will give you control over items in the list. In a nutshell,
OpenURL is a standard protocol for describing information resources.
Among other things, OpenURL tries to unambiguously denote an
information resource with authors, titles, volume numbers, years,
unique identifiers, etc. and sends that information back to a
resolver. The resolve then comes back and says, "I can provide you
with these services, usually getting the full text, against that
resource."

In short, standards are definitely the way to go. Start with standard
metadata formats. Store the data in a standards-based way in
databases. Index the data and provide access to the index via SRU.
Provide access to individual items through OpenURL. When it comes to
other aspects of the ILS exploit the ILL protocol. Exploit the NISO
thesaurus standard. Use EDI for billing. Etc. This is not rocket
science, just normal everyday computing.

--
Eric Lease Morgan

I'm hiring a Senior Programmer Analyst.
See http://dewey.library.nd.edu/morgan/programmer/.