[Moims-dai] Reminder of Skype call tomorrow (Tuesday)
david at giaretta.org
david at giaretta.org
Mon Mar 14 19:58:20 UTC 2022
Meeting: 10:00 Tuesday (Washington DC time), 1400 UK time, 1500 Paris time.
NOTE USA HAS CHANGED TO DAYLIGHT SAVING SO THE EUROPEAN TIME HAS MOVE AN
HOUR THIS WEEK. BACK TO NOTMAL NEXT WEEK.
*Don't have Skype yet? Download it before you join <https://www.skype.com>
Review of agenda
APOLOGIES FROM ME FOR THIS MEETING. I HOPE JOHN WILL START THE MEETING,
1. Work on OAIS-IF design, prototyping and book - Steve will provide a
a. See discussion below about some ideas for usability
b. The latest version of the draft GB
I've updated the proposed interfaces and associated javadocs for thee
OAIS-IF generic to generic communications. They can be seen at
There are associated changes to the OAIS programming interfaces
-summary.html to add functionality which is strongly implied in OAIS Ref
Model, particularly about the grouping of RepInfo.
By the way, it you click on the diagram in javadocs you will get the full
size diagram which is easier to read.Interactions - from Steve -
c. WHITE BOOK FROM Steve -
d. Steve may sent out updated document/diagram
Ideas for usability:
1.1 Negotiation for Usability
One of the key capabilities which an OAIS should be able to provide is
Representation Information to accompany the Data Object to make an
Information Object which can be understood by its Designated Community,
however the OAIS may have more Representation Information, perhaps kept in
anticipation of changes in Knowledge Bases. A non-OAIS archive may have some
RepInfo so here we focus on OAIS archives but remember that the archive
(whether OAIS or not) may return "null" if it cannot supply what is needed,
and this may mean that the Data Object cannot be made to be usable by the
For those outside the Designated Community the OAIS can get Representation
Information which may be enough for the Information to be enough.
Alternatively, it may be necessary to Transform the Data Object in order for
the Information to be usable by someone in this other group of users.
1.1.1 Getting enough RepInfo
Thinking in general about the one requesting information getting enough
RepInfo one can think of a number of steps:
1. Ask if it already has enough RepInfo e.g. checkUsability which would
a. "true" - which means that there is enough RepInfo so nothing more is
b. "false" - which means that either more RepInfo [SEE (2)] OR a
Transformation is needed [SEE (3)]
2. Options for getting more RepInfo:
a. Get everything that the archive has - e.g. the whole RepInfo Network
(RIN), at least as a network of pointers BUT
only using pointers then somehow the requestor must be able to recognise the
identifiers, which might be OK if they use the same registry
1. NOTE: if the Data Object is a FITS file then if the RepInfo is a PDF
of the FITS standard then that would be OK because one would see that the
RepInfo needed is a PDF reader, which I have.
2. if the RepInfo is a piece of DRB then what the requestor needs to
recognise is NOT that specific example of DRB but rather the requestor would
be looking for the DRB processor which is used to extract information from
the combination of the DRB file and the Data Object.
a. As an alternative to getting the full RIN then one could get the
nearest pieces of RepInfo i.e. the Structure, Semantic and Other RepInfo for
the Data Object (which might itself be the DO for some RepInfo).
b. If nothing will work then see (3)
2. Negotiate a Transformation by sending a list of acceptable formats
for the DO - but would need more details such as how Semantics are dealt
with. The archive would need to see if any of these formats are acceptable
e.g. in terms of Transformational Information Properties.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the MOIMS-DAI