[openbiblio-dev] Preliminary list

[Dan Sheppard]

Dear Jonathan,

For me, I think it is best to look at this from another perspective.

First identify your end users. These could be library readers or, in 
your case, perhaps various aspects of civil society, academics, and so 
on, as well as other libraries. Try to pin down their goals, and so on.

Then look at what they want from your data, one by one for each group, 
and why they are likely to need your metadata to access and process it. 
Examples of this might include

   * users will seek data, and the metadata can help them find it. What 
metadata will the users have in-hand? For example, while publisher 
location is traditional in bibliographies, and useful for 
disambiguation, how valuable is it here?

   * if your resource itself is difficult to get at, eg a book in a 
closed stack, you need a better surrogate for the data than if it is an 
internet resource, because the user will want to be more certain before 
going to the bother.

   * future users will rely on you having captured any metadata 
necessary for format migration should a format have become obsolete

   * to share with other libraries, they might require a certain set of 
fields to co-operate.

   * ...

That should lead you to a natural list of fields.

When that's done, I think you've reached the point where you can map it 
into various standards, such as these linked-data systems. A developer 
can easily remap them as a small part of the ingest process to target 
whatever system is being tageted. (There is no problem, with 
triple-based linked data, with adding fields to support exciting things 
which have not been anticipated by the software).

If you've analysed the requirements from an end user perspective, then 
you're on firm ground, whatever interchange and storage standards throw 
at you.

Dan.