[Moims-rac] Thoughts on RAC and Provenance

[Mark Conrad]

Don,
 
See my responses*** in the text of your message.
 
Mark
 
Mark Conrad
Electronic Records Archives (ERA)
NHER 
The National Archives and Records Administration
Building 494 Second Floor
310 State Route 956
Rocket Center, WV  26726
 
Phone: 304-726-7820
Fax: 304-726-7840
Email: mark.conrad at nara.gov 
ERA Website: http://www.archives.gov/era/

>>> Donald.M.Sawyer at nasa.gov 10/29/2007 9:21 AM >>>

Hi Mark,
I only have time to respond in writing to your first two items.  I
expect to be at the megameeting, but may be late.

cheers-
don


On Oct 26, 2007, at 1:26 PM, Mark Conrad wrote:
1. I strongly agree that , "...we need to discuss and agree on a more
definitive working scope statement."

So, do you think that if an repository can not obtain what I've called
'pre-ingest' provenance, it should be declared to have failed a
mandatory requirement?
 
***As I stated in my previous message:
 
***9. Do we get perfect provenance information for every collection
back to it's creation? Absolutely not! Should a repository try to gather
as much provenance data as possible? Absolutely! I like the idea of
Cal's statement about provenance. I believe with a little word smithing
his statement would be a good one to use in the RAC document.
 
***If a repository does not gather at least some provenance information
from prior to the time of ingest then you could easily have a situation
where, "a Producer could hand the repository a Word document and say
that it was produced by Julius Caesar and the repository would be
considered trustworthy if it preserved, migrated and disseminated this
document reliably."

 
 
2. I am troubled by these statements:
 
"It then follows that the objective is to identify a core set of
requirements that all repositories should be able to meet.  However this
is putting things a little too strongly.  It may be possible to have
some requirements that either are not relevant to some repositories, or
are sufficiently general that the actual evidence of meeting them very
much depends on the type of, or particular, repository.  TRAC has a
small number of these types of requirements.  Having such requirements
makes it more difficult to devise a scoring system that is both fair and
understandable.  Therefore it appears desirable to avoid too many of
such requirements.  Note that TRAC does not define a scoring system."
 
I think that it will be extremely difficult to come up with a scope
statement based on the above and have it be meaningful/useful to a broad
set of repositories. I believe that we should either have a document
that contains only core requirements that all repositories should meet,
or have a document that makes explicit which requirements are
core/mandatory for all repositories and which are not mandatory and
subject to local interpretation. I would prefer the former. I believe
this will result in a more useful document.



As I understand your paragraph, you prefer a standard that sets out a
core set of mandatory requirements that all repositories should meet. 
If that is the case, then that is the same as the first of the two
options I gave above.  You seem to be restating my options in your
response.  My statement about what TRAC says is NOT the same as my view
of what we SHOULD put in RAC.  Am I missing something here?
 
***I like the first option that you propose - only core required
elements - but the very next sentences say,  "However this is putting
things a little too strongly.  It may be possible to have some
requirements that either are not relevant to some repositories, or are
sufficiently general that the actual evidence of meeting them very much
depends on the type of, or particular, repository." It sounds to me as
if you are saying that having a document with only core mandatory
requirements is NOT what we should do. 
 
***As I stated in my previous message, 
 
***I believe the RAC should either contain ONLY mandatory requirements

 
***OR 
 
***mandatory requirements and "some requirements that either are not
relevant to some repositories, or are sufficiently general that the
actual evidence of meeting them very much depends on the type of, or
particular, repository" AND explicit text about which requirements fall
are mandatory and which are not.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ccsds.org/pipermail/moims-rac/attachments/20071029/aff753a3/attachment.htm