Notes from Megameeting 24th September 2007


BarbaraSierman Koninklijke Bibliotheek, Netherlands
BruceAmbacher UM
SimonLambert STFC
DonaldSawyer NASA GSFC
HelenTibbo UNC
JohnGarrett NASA
KatiaThomaz INPE, Brazil
MarkConrad NARA
RobertDowns Center for International Earth Science Information Network (CIESIN), U Columbia

All the discussion at this meeting was conducted by chat, so the following transcript of the meeting (with a few typos corrected) is complete.

Conclusion of the discussion: MarkConrad and HelenTibbo to continue studying options for definitions and approaches to authenticity and related concepts, and relate them to the TRAC document to see what modifications to that document would be required.

cclrc >> (All): Hello Mark & Helen - this is Simon - David can't make it today.
Mark Conrad >> (All): Hello.
Helen Tibbo >> (All): Hi Mark, just saw your message and have sent additional definitions of "authenticity" "integrity," and "reliability" to the list.
Helen Tibbo >> (All): Sorry I couldn't make it last week - the traveling thing makes it hard to connect. 
RobertDowns >> (All): Hi all, I do not have a headset today, so only will be using chat to communicate.
cclrc >> (All): We seem to have been using chat exclusively recently - it means that the records of the discussions are complete.
Mark Conrad >> (All): Bruce, If you are saying something, I can't hear it
BruceAmbacher >> (All): Very low volume today.  So I will be just typing.
Helen Tibbo >> (All): Mark, nice internal analysis of the use of "authenticity" within the TRAC document
cclrc >> (All): Mark, is the key part of your email the final section "Based on its usage in the TRAC document one could deduce:"?
Helen Tibbo >> (All): From what I found looking at definitions of authenticity, there doesn't seem to be great disagreement (of course, 90% are from InterPARES folks), but there are wording differences.
Helen Tibbo >> (All): For UBC and InterPARES, integrity is the combination of authenticity and reliability
Katia Thomaz >> (All): hi. sorry to be late.
Mark Conrad >> (All): I am beginning to analyze some of the documents from the Preservation working group of InterPARES 1. They seem to have the most useful information on reproducing an authentic copy of an original.
Mark Conrad >> (All): Helen, Authenticity is defined by the InterPARES folks in terms of identity and integrity.
Mark Conrad >> (All): I will have more to report next week.
Mark Conrad >> (All): Where do we go in the mean time?
Helen Tibbo >> (All): Well, there is come circularity in the use of terms in interpares... see the integrity quotes I sent.
cclrc >> (All): Should we give Mark another week and meanwhile go on to B5?  Or does anyone have some comments about authenticity right now?
BruceAmbacher >> (All): Did anyone get definitions I sent from ISO 15489?  They seem to get hung up in MOIMS.
Helen Tibbo >> (All): no I did not get them
Mark Conrad >> (All): no.
Katia Thomaz >> (All): no I didn't
BarbaraSierman >> (All): no I did not
RobertDowns >> (All): no, I believe that I did not. Did you send them recently?
BruceAmbacher >> (All): cclrc: How do I rectify this?  I sent them last Monday and Tuesday via MOIMS, where they seem to get held for review.
cclrc >> (All): I don't remember seeing anything but I will check.  When you say review, who would be doing the reviewing?
Mark Conrad >> (All): What do you mean sent them via MOIMS?
Mark Conrad >> (All): Can you try sending the definitions in a reply to my message or Helen's?
JohnGarrett >> (All): I haven't seen anything that needed review
BruceAmbacher >> (All): I will address this with John offline.
Katia Thomaz >> (All): remember that authenticate is very different from authenticity
Mark Conrad >> (All): Authenticate is used in two different ways in the TRAC document. One way is directly related to authenticity.
Katia Thomaz >> (All): where?
Mark Conrad >> (All): The TRAC document uses the term "authenticate" in two different ways: 1. To authenticate a user of the archive (i.e., to verify the credentials and authorizations of a person attempting to use the repository). 2. To verify the authenticty of an object in the repository.
Katia Thomaz >> (All): i see autheticity in B6.10,and C1.6
Katia Thomaz >> (All): the others authenticate or authentication
RobertDowns >> (All): Definition of Authenticity from ISO/CD 15489-1:"An authentic record is one that is proven both to be what it purports to be and to have been created or sent bythe person who purports to have created or sent it."
Katia Thomaz >> (All): again, authentic or authenticity in B6,10 and C1.6. authenticate or authenication in the others.
Mark Conrad >> (All): See authentication on page 42 of the TRAC document.
Mark Conrad >> (All): B 6.10
Katia Thomaz >> (All): yes, item B6.10
Mark Conrad >> (All): Authentication is used at B. 6.4 and elsewhere with a very different meaning
Katia Thomaz >> (All): i think C1.6 must refer only to integrity, not authenticity
Katia Thomaz >> (All): because, as we discussed last week, repository only keep the integrity.
Katia Thomaz >> (All): or not?
Mark Conrad >> (All): No. A repository has to be able to demonstrate that it can reproduce authentic copies of the originals it ingested.
Katia Thomaz >> (All): ok, only for the COPIES
Mark Conrad >> (All): A repository can only disseminate copies.
BruceAmbacher >> (All): NO.  For all versions
Katia Thomaz >> (All): not the digital object
BruceAmbacher >> (All): How can a copy be authentic if the digital object it is created from is not authentic and the chain of custody and use demonstrates that?
Katia Thomaz >> (All): because only the producer can evaluate the authenticity of the digital object
Katia Thomaz >> (All): judge is better than evaluate
RobertDowns >> (All): The evidence should be that the object was received from the producer or the representative of the producer.
Mark Conrad >> (All): No. Assessments of authenticity are made by other than the creator - sometimes long after the creator is dead.
BruceAmbacher >> (All): Yes.  But the repository must demonstrate that it has continued to keep the nformation in the object what it was when received.  That is the role of audit trails, etc.
Katia Thomaz >> (All): so the integrity
Mark Conrad >> (All): Read B. 6.10
Mark Conrad >> (All): As Ken Thibodeau would say, You cannot preserve an authentic electronic record - only the ability to reproduce one.
BruceAmbacher >> (All): The best course here seems to be to use established definitions from national/international groups such as ISO 15489, InterPARES, SAA, current TRAC language?  I understand we all have to get our minds around some subtle differences but much of this was debated in developing these sections in TRAC.
Katia Thomaz >> (All): i agree with bruce
BruceAmbacher >> (All): Mark Would Dr. T. accept the concept of authentic information (digital object)?
Mark Conrad >> (All): He would say that you can store the bits and procedures to reconstitute an authentic digital object.
Mark Conrad >> (All): There is a 25-page paper on this in the report from InterPARES 1.
BruceAmbacher >> (All): And work done by Lockheed for ERA?
Katia Thomaz >> (All): but this is another aspect
Don Sawyer >> (All): Does anyone have a problem with the definition that Robert gave earlier, as follows:  Definition of Authenticity from ISO/CD 15489-1:"An authentic record is one that is proven both to be what it purports to be and to have been created or sent bythe person who purports to have created or sent it."  It works for me.
Katia Thomaz >> (All): i must compare with the others to reach a conclusion
Helen Tibbo >> (All): I think the task here is to agree upon definitions that clarify our work and clarify the ulitmate standard for the end user. I don't think we should be devising new definitions, but we need to have some clearly stated that keep us clear on what we mean throughout the document.
BruceAmbacher >> (All): ISO 15489-1 also has a definition of Integrity.  Both work for me, although their primary focus is records management, not preservation.
Mark Conrad >> (All): Bruce, I don't know what Lockheed Martin is using as an operational definintion. There is a definition in the ERA requirements document.
Don Sawyer >> (All): When OAIS talks about disseminating copies of the originals, the 'originals' referrs to what the Repository received.
Katia Thomaz >> (All): yes
BruceAmbacher >> (All): Which are presumed to be authentic
Don Sawyer >> (All): Yes, assumed to be so by the repository/archive.
Mark Conrad >> (All): TRAC says that for some communities the chain of custody has to extend back beyond ingest.
BruceAmbacher >> (All): Which is all any repository can do, then follow procedures to maintain that object or show how it is modified to keep it viable.
cclrc >> (All): So an original arrives with a presumed authenticity, and the repository must maintain the status of its being what it purports to be.
BruceAmbacher >> (All): The submission agreement and the standard procedures between the producer and the repository should address that to the satisfaction of the designated community - and it all should be available to users.
Katia Thomaz >> (All): what about repositories that get digtal  objects from the Internet without negotation with the producer?
Mark Conrad >> (All): In that case the repository is the collector or the producer.
BruceAmbacher >> (All): They can only be maintained "as received"  Their provenance is documented as to their capture.
Katia Thomaz >> (All): and so...
Katia Thomaz >> (All): what about authenticity in these cases?
Helen Tibbo >> (All): I think key in the collecting scenario is that the user understands the source/collection methodology and its limitations
Mark Conrad >> (All): The collected information is an authentic collection made or received by the repository in the course of its business.
BruceAmbacher >> (All): The documentation that defines capture procedures should address Helen's point by stating what those procedures are and the degree of authenticity" that can be achieved.
Katia Thomaz >> (All): i think this point must be clear in the document
BruceAmbacher >> (All): So, how have we advanced our effort today?  Where are we?  What is the next step?
Katia Thomaz >> (All): there are two kinds of repositories: one that have an producer and one that haven't
Mark Conrad >> (All): No. In your example the producer and the repository are the same person.
Don Sawyer >> (All): I would like to see some agreed definitions.  Then, we can address how to interpret in various situations, such as internet capture.
Katia Thomaz >> (All): ok.
cclrc >> (All): Could we take the ISO 15489-1 definitions and task someone to relate them to the TRAC doc to see if any problems arise?
Helen Tibbo >> (All): While we have talked about authenticity, I am not sure we have come to grips with related terms and we haven't discussed "understandability" at all. DRAMBORA talks a lot about understandability alongside authenticity.
Mark Conrad >> (All): Those are two separate concepts.
Katia Thomaz >> (All): i agree
BruceAmbacher >> (All): Definitely.
Mark Conrad >> (All): Simon, I will look at the ISO definitions as part of my work on authenticity.
Helen Tibbo >> (All): Yes, they are, but hopefully a repository can provide both authentic and understandable documents.
cclrc >> (All): Thanks Mark.
Helen Tibbo >> (All): Mark, how can I assist?
Don Sawyer >> (All): OAIS takes understandability to be very important and there has been a lot of ciiscussion about it in arriving at TRAC.  This does not mean TRAC is perfect, of course!
Mark Conrad >> (All): I believe that if we adopt the ISO definition we will have to make some changes to the TRAC document.
Mark Conrad >> (All): I will be sure to include that in my analysis.
JohnGarrett >> (All): I think I would prefer making some changes in TRAC document rather than coming up with a new definition
Katia Thomaz >> (All): ok. glossary is the focus now.
BruceAmbacher >> (All): But a records management definition may not be the best from a preservation perspective.
BruceAmbacher >> (All): We need to examine any/all
Katia Thomaz >> (All): i think ISO 15489 refers to management and preservation
Mark Conrad >> (All): I agree with Bruce.
Helen Tibbo >> (All): I will look through the TRAC document regarding understandability and look for external definitions. 
cclrc >> (All): Perfaps this is too optimistic, but maybe there is one definition that matches better than others with TRAC and would require fewer modifications.
Mark Conrad >> (All): That is part of what I was looking for this past week.
Katia Thomaz >> (All): folks, i must go now.
cclrc >> (All): Looks like we have actions on Mark and Helen to come back to us with results/recommendations for discussion
cclrc >> (All): based on existing definitions and their implications for TRAC
Katia Thomaz >> (All): i'll follow the wiki and emails
Don Sawyer >> (All): O'K, works for me.
Don Sawyer >> (All): Bye
Katia Thomaz >> (All): bye and have a nice week
Helen Tibbo >> (All): Bye and Happy Fall!
cclrc >> (All): Bye everyone - record will go on the wiki as usual.
Mark Conrad >> (All): bye!
JohnGarrett >> (All): bye

-- SimonLambert - 24 Sep 2007

Edit | Attach | Watch | Print version | History: r2 < r1 | Backlinks | Raw View | Raw edit | More topic actions
Topic revision: r2 - 2008-02-13 - KatiaThomaz
This site is powered by the TWiki collaboration platform Powered by PerlCopyright © 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback