data sharing

[Peter Rauch]

On Thu, 3 Dec 1998, Hugh Wilson wrote:
> The 'citation' problem requires some sort of solution, maybe via firm
> 'versioning' of old datasets to an archive or, perhaps, by including
> some sort of record-specific 'edit trail' that would allow the user
> to track change.

I think provision of audit trails are a _critically_ important issue in
the design and implementation of collections databases.

>  But, if 'value' relates to public usage, then
> priorities should be focused on the broader (non scientific)
> potential user base which, I think, could care less about reference
> citation.

Can you explain your comment, please. In particular, I think whether a
non-scientific (public?) user cares less or more about audit trails
(edit trails?, reference citations?) will more reflect that user's [lack
of] understanding of the problems with using non-data (i.e., data that
were once believed to be accurate, and are now known to not be the
[correct] data).

"Value", as a criterion of a collection, should not derive from "public
usage", but from the benefits accrued from that public usage. It doesn't
matter how much the data are used (by the public or by science) if the
data lead to bad policy, bad decisions, results based on fictitious
data.

Who are these "non-scientific" users --users who don't need to care
about whether the data they used are real or fictitious? Should these
users be the concern of (be serviced by) the museum collection manager?
Should the museum collection manager design her information system to
provide (some sort of) data to people who could care less whether the
data are real or fictitious?

Perhaps a collection can finesse its responsibility to provide an audit
trail by stating to each person who accesses its database "Beware: the
data you are about to use may contain errors; any use to which you put
these data will thus be suspect to the degree that it is sensitive to
the use of defective data. This collection will not be responsible for
any conclusions or uses based on our data."

Even a collection which institutes a super audit trail system should
make the above disclaimer. But, at least this latter will be able to add
"You may revisit our database at any time, to determine whether any of
the data you used has changed."

But, maybe I didn't understand your point. So, please, elaborate.
Peter