[Web4lib] Metadata tools that scale

[Mark Phillips]

At the University of North Texas we have built a generic tool for
metadata creation which we use for all of the projects that we take on
in the Digital Projects Unit.  The tool is a web form based around a
locally qualified Dublin Core element set with input rules that are well
documented but still customizable for different projects.  The system
allows for making default "templates" which allow the different projects
to have common values preloaded when they visit the webform.  The
metadata tool makes sure that required fields (we require several fields
for our systems) are present and then allows the student to submit the
form.  There is an XML file created which is then placed along with the
files we are digitizing and they ride together for the rest of the
workflow and into our systems. 

We have found that these forms are simple to use, and with proper
documentation for each project,  it is quite easy to get our student
works to create high quality metadata for different collections. 

On the flip side of things,  when we get content into our systems we
have another tool which allows us to look at all of the metadata and
find mis-spellings, typos, and makes it easy to create authoritative
versions of names and normalize all of our data in order to provide a
better search and browse system for our users.  

So far these two systems are things we coded up in php with a little
mysql here and there,  but I think in the next few months we will be
moving them to one of the popular web frameworks if we can get some free
time to  do it. 

Just an idea of how we do the metadata stuff.


Mark









Mark Phillips
Digital Projects Unit Manager
Digital Projects Unit
University of North Texas Libraries
P.O. Box 305190
Denton, TX 76203-5190
        (940) 369-7809
Fax: (940) 565-2599
http://www.library.unt.edu/digitalprojects/
>>> Will Sexton <will.sexton at duke.edu> 01/23/07 1:59 PM >>>
I'm participating in a group that's looking at the digitization program 
here at Duke U. Libraries.  The topic brings with it a wide range of 
questions, but one that we return to often relates to the lack of a 
suitable metadata tool.  In the past, our approach to creating metadata 
has mostly involved one-off development in platforms like FileMaker and 
Office (Access, Excel).  Experience has made me extremely averse to this

approach, and in my judgment, to have a digitization program that scales

at all, we need much better support for metadata creation.  A suitable 
platform would have the following features:

* authority control
* adaptability to diverse metadata schemas
* can be integrated with:
  - Digital Asset Management, for storing both metadata and files 
resulting from digitization
  - some sort of Content Management (i.e., for Web publication and other

modes of dissemination)
  - a tool for managing and tracking the digitization process, itself a 
highly theoretical entity
* is web-based for distributed access
* provides for workflow management

We ask ourselves, for these requirements, should we build or buy?  But 
is there even anything to buy?  How do others involved in supporting 
digitization (and metadata creation, generally) deal with this issue?  
Finally, does anyone have any wisdom to share which would contravene the

features list or other assumptions stated above?

My sense of this question is that the options for "buy" (or "borrow", or

"steal", though "beg" may have some potential) are very limited.  But 
it's one point on which I'd be very happy to be proved wrong.

Best,

Will

-- 
Will Sexton
Metadata Analyst / Programmer
Duke University Libraries

_______________________________________________
Web4lib mailing list
Web4lib at webjunction.org
http://lists.webjunction.org/web4lib/