FRBR

39 views
Skip to first unread message

Jerome

unread,
Mar 26, 2010, 11:55:25 AM3/26/10
to OAI-ORE
I have a question that has arisen as a result of some of our work
trying to create AIPs for antique games using OAI-ORE. For a variety
of reasons (most of them having to do with access), it's very nice to
be able to record metadata about games using FRBR Group I entities as
a model. However, in terms of the realities of the actual data files
I'm collecting for the games, I've got single copies (which is to say,
in FRBR terms, I've got an Item).

Given the 1:M relationships between FRBR group 1 entities, if I want
to completely model a FRBR hierarchy for a work in OAI-ORE, I need at
least 4 resource maps: 1 work-level resource map to aggregate
expressions, an expression-level map for aggregating manifestations, a
manifestation-level resource map to aggregate items and at least 1
item-level resource map to aggregate the actual content for a
particular item.

For games where there are significant issues around editions (and
there's a lot of those in the world), this actually works as a pretty
good way to model things, making it easy to add new resource maps for
particular editions/revisions, new copies of items, etc. as they come
along. But there are some games where the situation is much simpler,
and there's not a lot of reason to have work/expression/manifestation
resource maps around, because all I have is one item and that's all
I'm likely to ever have.

In those cases, how evil would it be of me to have an item-level
resource map indicate that it is aggregated by manifestation/
expression/work aggregations that actually don't have resource maps?
that is to say, I would have references in the item-level resource map
to URIs for work/expression/manifestation aggregations, and use the
ore:aggregates verb to show relationships between them, but I wouldn't
actually record the existence of resource maps for those upper-levels
of the FRBR hierarchy. I *would* probably associate other metadata
with those higher level FRBR aggregation URIs (e.g., the Work-level
URI would be the subject of a mods:title predicate with a string as
the object). Doing that would mean I'd only need a single resource
map to describe a whole FRBR hierarchy in instances where I don't need
to reference actual content files for multiple items/manifestations/
expressions.

I'm not convinced this is necessarily a good idea anyway, since
there's some argument to be made that consistency of implementation of
FRBR hierarchies in OAI-ORE is preferable, but I'm wondering if people
*wanted* a simple, one resource map approach, this might be acceptable.

Robert Sanderson

unread,
Mar 26, 2010, 2:59:10 PM3/26/10
to oai...@googlegroups.com

Hi Jerry,

On Fri, Mar 26, 2010 at 8:55 AM, Jerome <jmcd...@uiuc.edu> wrote:
I have a question that has arisen as a result of some of our work
trying to create AIPs for antique games using OAI-ORE.  
 
  But there are some games where the situation is much simpler,

and there's not a lot of reason to have work/expression/manifestation
resource maps around, because all I have is one item and that's all
I'm likely to ever have.
In those cases, how evil would it be of me to have an item-level
resource map indicate that it is aggregated by manifestation/
expression/work aggregations that actually don't have resource maps?


To make sure that I understand ...

Rem-Item ore:describes Aggr-Item
Aggr-Item ore:aggregates ".../wizardryv.zip"     // etc.

Aggr-Item ore:isAggregatedBy Aggr-Manifestation
Aggr-Manifestation ore:isAggregatedBy Aggr-Expression
Aggr-Expression ore:isAggregatedBy Aggr-Work
Aggr-Work dc:title "Wizardry V" // etc.

And not say that Aggr-Work, -Expression or -Manifestation are ore:describedBy anything.

What you wouldn't want to do is the other way around, eg include all of the information into the Rem-Work, otherwise you'd be falling foul of the one aggregation per resource map rule.

I'm not convinced that this is evil at all.

Rob

Jerome

unread,
Mar 26, 2010, 3:21:58 PM3/26/10
to OAI-ORE
You have it exactly right. Pushing the idea a little further, I'm
imagining a system where, if I only have a single item to worry about,
I create a single resource map for the Aggr-Item; URIs are assigned
for the Aggr-Manifestation, Aggr-Expression and Aggr-Work, but they
have no resource map. Or they don't, UNTIL I add another copy of the
item, at which point I add a new resource map for the 2nd Aggr-Item,
and a new resource map for the Aggr-Manifestation which aggregates the
two Aggr-Items. If I add a different Aggr-Manifestation (say, the web
distribution of the game rather than the CD distribution), then I add
another Aggr-Manifestation resource map, AND an Aggr-Expression
resource map that aggregates the two Aggr-Manifestation. Basically,
you have a virtual FRBR tree which you populate with resource maps
from the item-level up as you add new items/manifestations/
expressions.

On Mar 26, 1:59 pm, Robert Sanderson <azarot...@gmail.com> wrote:
> Hi Jerry,
>

Reply all
Reply to author
Forward
0 new messages