Notes from Megameeting 22nd December 2008

Attendees:

BruceAmbacher UM
DavidGiaretta STFC
HelenTibbo UNC
JohnGarrett GSFC
MarieWaltz Center for Research Libraries
MarkConrad NARA
RobertDowns CIESIN, Columbia University
SimonLambert STFC

The discussion focussed on the different senses of the word "access" in section B6 and the implications. The agreed approach was to:

  • update the glossary for Access Policy
  • identify which metric needs a restriction of access i.e. only one type of access
  • add in examples and/or discussion for several types of access
  • identify if anything is missing from the B6 requirements taken as a whole

The next meeting will be in two weeks, 5th January.

Actions:

  • MarkConrad to redraft section B6 for discussion.
BruceAmbacher >> (All): Has anyone responded to Mark's email about whether the 
NARA meeting room will be satisfactory?  It is a mini auditorium with a pc and 
screen.
Mark Conrad >> (All): No
BruceAmbacher >> (All): No - no one has responded or No my description is wrong?
cclrc >> (All): oops - anything with a screen and a way to display would be 
fine I guess
Mark Conrad >> (All): No one has responded other than you Bruce,
JohnGarrett >> (All): Hi, I guess I missed the question.   Sounds OK to me.
RobertDowns >> (All): Sounds ok to me, too.
Mark Conrad >> (All): The seating arrangement is not ideal for people using 
their own computers, but it will be good for group editing of the document with 
one recorder.
Mark Conrad >> (All): There is no wireless available in the room.
BruceAmbacher >> (All): Mark, Can you arrange for a second connected pc?
cclrc >> (All): That might be a plus
Mark Conrad >> (All): We can probably provide folks with access to the internet 
in the ERA Research lab on the breaks.
cclrc >> (All): Sorry - no wireless may be a plus
JohnGarrett >> (All): Well maybe then people will have to participate in the 
meeting rather than on their email
Mark Conrad >> (All): Bruce, I don't think there is a way to have a second 
computer connected to the projector. What did you have in mind for that second 
computer?
BruceAmbacher >> (All): random internet access, other documents displayed.
BruceAmbacher >> (All): And what is our starting point today? B6.4?
Mark Conrad >> (All): I will see what can be done, but if I remember correctly, 
the connectivity to the internet and projector is fairly constrained.
Mark Conrad >> (All): I took an action item at the last meeting to try to 
redraft all of B.6. I did not complete that draft.
Mark Conrad >> (All): See my comment at the beginning of B.6.
Mark Conrad >> (All): Neither the TRAC document or the OAIS are good at making 
it clear which sense of the word access is being used at any particular point in 
the text.
Mark Conrad >> (All): We will need to make that explicit for each requirement.
cclrc >> (All): In OAIS capitalisation indicates the functional entity but 
otherwise it is take as "normal" meanings
Mark Conrad >> (All): I have listed four of the different meanings of access 
that can be found in the TRAC and OAIS documents. 
JohnGarrett >> (All): Yes, I think we should make clear what sense of access we 
are checking in the requirements.
Mark Conrad >> (All): John, We will need to check multiple meanings of the word 
access.
cclrc >> (All): Is this throughout the docuemnt or just section B6?
Mark Conrad >> (All): I found some references to access in Section A. as well.
BruceAmbacher >> (All): And how many of the four uses of sccess are capitalized 
to capture an OAIS specific use?
Mark Conrad >> (All): Bruce, All four meanings and more are found throughout 
OAIS  capitalized and not.
Mark Conrad >> (All): There are 316 occurrences of access in the OAIS document.
BruceAmbacher >> (All): Yes, but as David indicates any that are capitalized 
should conform with an OAIS defined use.
RobertDowns >> (All): One way to differentiate these might be to system access, 
catalog description access, DIP access, and restricted access.
BruceAmbacher >> (All): Robert, good solution - always require a reference to 
the type of access.
cclrc >> (All): Robert - sounds good, but do we only have to put in extra words 
if the meaning is unclear
cclrc >> (All): ...I assume that in many cases all types of access are meant
RobertDowns >> (All): Yes, if we are referring to all definitions of access in 
the general sense, we might not want to qualify the term.
JohnGarrett >> (All): Even the capitalized form in OAIS will refer to the Access 
function, which could include several of the meanings of access in its 
subfunctions.
Mark Conrad >> (All): David, How do you refer to all types of access in one 
requirement?
Mark Conrad >> (All): I would invite you all to do a search in the OAIS on the 
term access before making these generalizations.
RobertDowns >> (All): I agree with the implication that, if possible, we should 
try to be as specific as possible to reduce ambiguity.
cclrc >> (All): Hmm - maybe you're right. I was thinking that accessing a part 
of a record implies all the other types of access
Mark Conrad >> (All): For example, in some places the OAIS uses the term access 
to refer to discovering an AIP and receiving a DIP as access. At other points 
the OAIS draws a distinction between these two activities calling the former 
access and the latter dissemination.
cclrc >> (All): Mark - looking at B6 there are references to "access policies" 
which presumably covers all types of access
Helen Tibbo >> (All): Hello! Sorry to just be a lurker today - had a Skype 
project call and now I need to run to another meeting.
Helen Tibbo >> (All): Give me an assignment
Mark Conrad >> (All): Rewrite B. 6. ;)
Helen Tibbo >> (All): That could produce interesting results!
BruceAmbacher >> (All): Helen, was Mark's wink harrassment?  I will back you up!
cclrc >> (All): Also "access requests" could refer to all types of access.
Mark Conrad >> (All): David, You would need policies for all of the different 
meanings of access. If we do not make that explicit and list the types of access 
the policies should cover we have not made good requirements.
Mark Conrad >> (All): Access requests are divided into at least three separate 
type of requests in the OAIS.
Mark Conrad >> (All): Three categories of Consumer requests are distinguished: 
query requests, which are executed in Data Management and return immediate 
result sets for presentation to the user; report requests, which may require a 
number of queries and produce formatted reports for delivery to the Consumer; 
and orders, which may access either or both Data Management and Archival Storage 
to prepare a formal Dissemination Information Package (DIP) for on- or off-line 
delivery. pp. 4-15 and 4-16.
cclrc >> (All): Mark - I suppose we should identify where the term access is 
specific to one or other of the types and make it explicit and also define the 
various types of access - maybe in the glossary? 
JohnGarrett >> (All): Maybe, we don't need requirements for every sense of 
access.  I suggest we go through the ones here and decide if we need to qualify 
the type(s) of access for that requirement.   Then at end look over the whole 
list and see if we can identify any requirements that we want to still include.
Mark Conrad >> (All): I think we will need to make it explicit which meaning of 
the word access we are using each time the word appears in the document. I like 
Robert's approach of adding a modifier. We would need to put a glossary item for 
access with each of its modifiers.
Mark Conrad >> (All): John, Shall we start with B.6.1.?
cclrc >> (All): SOunds good - let's be explicit
Mark Conrad >> (All): I think B.6.1. will do a great job of explaining the 
difficulties. Which of the four or meanings of access do you think apply for 
B.6.1.?
JohnGarrett >> (All): Yes, let's start with B6.1.  But I think the list will 
start of the types of access will start getting longer than the requirement.
cclrc >> (All): " access and delivery options"  - I think could include all 
meanings of access
Mark Conrad >> (All): John, I think it makes more sense to have separate 
requirements for each meaning of the word.
JohnGarrett >> (All): I think B6.1 applies to all 4 of Mark's meanings
Mark Conrad >> (All): And so how do we explain that to the readers of this 
document?
JohnGarrett >> (All): I don't think that is reasonable.  I think we'll have a 
100 requirements in B6.
cclrc >> (All): We could define the various aspects of access and say that 
unless otherwise specified the terms includes all these types
cclrc >> (All): The question then is whether one or more of the types needs 
special treatment
Mark Conrad >> (All): David, The supporting text, evidence, and discussion 
sections would have to contain different information for each meaning of the 
term access intended for that requirement.
cclrc >> (All): Mark - OK - I can see that we could provide different examples 
and discussion for each meaning
Mark Conrad >> (All): In other words, what you would check for would be 
different for each meaning of the term.
cclrc >> (All): .... I thought you were talking about the metrics themselves 
cclrc >> (All): .....which is where I was having difficulties
RobertDowns >> (All): Since some repositories might be dark, with limited access 
services, we might want to state that the policies should appropriate to the 
types of access services offered by the repository. 
Mark Conrad >> (All): If you have a metric without some explanation of what to 
check for in the supporting text do you have a useful metric?
BruceAmbacher >> (All): Perhaps it would be clearer if we rewrote one 
requirement to illustrate.
cclrc >> (All): Mark - I see what you mean - makes sense. We are in any case 
very short of examples
Mark Conrad >> (All): Bruce, Which one would you like to use as the guinea pig?
Helen Tibbo >> (All): Sorry to have to run - never thought this week would be so 
busy. Happy Holidays to all!!!!
cclrc >> (All): ....and as Robert points out there are some special cases we 
could highlight
cclrc >> (All): Happy Holidays Helen
BruceAmbacher >> (All): B6.1 and B6.2 already have "modifiers" for their use of 
access.  B6.3 will require several modifiers to define all types of access 
implied - as I see it.  Do others agree/disagree?
cclrc >> (All): Access policies seem to appear several times in the metrics as 
well as the discussion
Mark Conrad >> (All): I do not see a modifier to access in B.6.1.
cclrc >> (All): Bruce - I don't see the modifiers either
BruceAmbacher >> (All): B6.1 is modified by the fact that the repository will 
document what access policies it has.  So if one asks the repository will 
provide one/many responses/policies to meet the query.
Mark Conrad >> (All): Don't we need to specify the types of access policies that 
the repository must have?
cclrc >> (All): Ahha - not what I would have called a modifier. 
Mark Conrad >> (All): Bruce, What if the policy the repository produces only 
covers one meaning of the term access?
cclrc >> (All): We could add some text to Access Policiy in the glossary - to 
give examples of the types of policies, and also add some examples of access 
policies in the examples
Mark Conrad >> (All): Examples are not mandatory. Need something in the 
supporting text for each type of access.
BruceAmbacher >> (All): The requirement says the repository will document what 
access policies it has.  So if it is dark it would not have a user access to 
content; if it is open it would define what user can see/do/request.
cclrc >> (All): Bruce - so we need to put in guideance as to what would be 
adequate
BruceAmbacher >> (All): Sorry, I cannot type fast enough in real time
Mark Conrad >> (All): By definition, an OAIS makes its AIPs visible and 
available to its Designated Communities. Section 3.2.6., pg. 3-5
RobertDowns >> (All): So a dark archive would offer restricted access.
JohnGarrett >> (All): I think we could add some discussion of the different 
parts/types of access policy either in the introduction to B6 or in the 
discussion of B6.1.  Then it would be up to the auditor to determine if the 
evidence was adequate.
Mark Conrad >> (All): I am not sure that a dark archive would fit the definition 
of an OAIS unless the designated community is defined as those who have access 
to the dark archive.
BruceAmbacher >> (All): This discussion raises a larger question of whether TRAC 
was intended to be a fully compliant son of OAIS.  It was developed in the 
spirit and framework of OAIS but not fully compliant in language etc.  Do you 
agree David?
JohnGarrett >> (All): The designated community may be extremely limited.
JohnGarrett >> (All): I would like it to be fully compliant.
Mark Conrad >> (All): John, So you would not make having a policy that related 
to each type of access mandatory?
BruceAmbacher >> (All): Dark archives designated community is those with right 
of access and the need to know
Mark Conrad >> (All): I thought the purpose of this document was to design 
metrics for testing OAIS compliance.
cclrc >> (All): Bruce - I don't think there is a problem - OAIS terms are pretty 
flexible
cclrc >> (All): Yes, I think we need OAIS compliance
JohnGarrett >> (All): No, I don't think it is absolutely mandatory for example 
for a repository to say we will only disseminate in these particular formats.  
It might be a good idea, but not necessarily mandatory.
BruceAmbacher >> (All): My concept of the purpose of TRAC, which is subject to 
modification by this group as it works through document, was to develop an audit 
and certification requirements program, acceptable to the international 
community.
Mark Conrad >> (All): I thought one of the objectives was to provide metrics 
that claims of OAIS compliance could be checked against.
JohnGarrett >> (All): Bruce, do you think being compliant with OAIS would make 
it unacceptable to the international community?  I think making it compliant 
with OAIS makes it more likely that it will be acceptable.
BruceAmbacher >> (All): John, I agree.  But TRAC was not in blind lockstep with 
OAIS.  Our document can be where/if needed.
BruceAmbacher >> (All): OAIS is not infallible as the 5 year review is 
demonstrating.
cclrc >> (All): Yes, I think it is a plus to be OAIS compliant but compliance is 
pretty narrowly defined. We can also use OAIS terminology where useful, and OAIS 
terms are pretty broad so I don't see a problem
Mark Conrad >> (All): John, Which meanings of the word do you think are 
mandatory for B.6.1.?
JohnGarrett >> (All): I would probably would not change the requirement 
statement, but I would add some additional text to the discussion.
cclrc >> (All): Bruce - I guess we have time to make OAIS and this document 
compatible - updating OAIS if necessary
Mark Conrad >> (All): Disscussion is not mandatory. Are you saying there are no 
types of access policies that are mandatory? 
Mark Conrad >> (All):  I am still willing to try to redraft this section, but I 
need some guidance on how to proceed. I do not sense that there is consensus on 
how to proceed.
JohnGarrett >> (All): Perhaps something like "The repository should include in 
its access policy information about who should be able to make use of the 
repository, what repository records (or parts of records) are available to which 
repository communities.  The repository should also make known what 
dissemination options are available (e.g. what formats, what media, associated 
costs).
BruceAmbacher >> (All): Lets be careful here.  B6.1, like other sections (B1, 
B2,  etc.) sets up the more specific requirements in the subsequent parts of B6. 
 We have had problems when we try to shove too much into one requirement without 
looking at the other requirements in that area.
Mark Conrad >> (All): Bruce what is your suggestion?
cclrc >> (All): I suggest (1) update the glossary for Access Policy (2) identify 
which metric needs a restriction of access i.e. only one type of access and (3) 
add in examples and/or discussion for several types of access - we may not need 
to duplicate examples of access policies
BruceAmbacher >> (All): Our efforts should be in the other B6 requirements.  Do 
they, together, develop all aspects of access we want to cover?  If not, what 
parts are missing?  Can we develop additional requirements or modify those that 
exist?
cclrc >> (All): Bruce - yes that should be (4) what is missing?
Mark Conrad >> (All): Anyone else want to meet next Monday?
Marie Waltz >> (All): not really
cclrc >> (All): SOunds like we have the week off
Mark Conrad >> (All): Ok. I will try to have something ready for two weeks from 
now for our discussion.

-- SimonLambert - 22 Dec 2008

Topic revision: r1 - 2008-12-22 - SimonLambert
 
This site is powered by the TWiki collaboration platform Powered by PerlCopyright © 2008-2018 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback