[Esip-preserve] A Minor Note on Nomenclature

Curt Tilmes Curt.Tilmes at nasa.gov
Tue Aug 24 17:39:46 EDT 2010


On 08/24/2010 04:54 PM, alicebarkstrom at verizon.net wrote:
> I think I'd be inclined to put "audit reports" in Context
> Information basically because they would not be used in the
> production scripts and are therefore produced by a special process.
> In the sense you're suggesting, it seems to me that an "audit
> report" is a verification of maintenance of proper record keeping of
> the production history. In my recent paper, I would be inclined to
> incorporate an "audit report" as part of "custodianship" - noting
> that the data structure for "custodianship" is not the same as the
> data structure for "production history". In the data structure in
> the paper, "custodianship" is represented by a graph that records
> something like who was on duty when, as well as the proper transfer
> of custodianship from one person to another. The mental model for
> this is a bit like a log book for night watchmen - although the
> model should probably be made more formal.  Good note. We do need to
> find a place for "audit reports".

Then we still disagree.  I think custodianship (and audit information)
is a critical component of "provenance" and not distinct from it.

production history says you used file X.

provenance includes where you got file X, when you got it, who was
responsible for it, who wrote the software that produced file X, what
compiler was used to compile the software, etc.

context I think of as a paper describing the format of file X, a
validation paper describing the usefulness of the data in file X, etc.

We need to make a big list of all the artifacts and map them to
those terms.

The whole set we can call "provenance and context" (we've done that
before).

Curt


More information about the Esip-preserve mailing list