All,
I've been following the fascinating discussions over the last week, and
have a few further comments in response to Bryan Lawrence's last
posting.
Bryan wrote:
> > Given that the development of preservation technology is at such an
> > early stage, I'm not sure anyone can answer that question easily.
What
> > we will be able to say with confidence that we can preserve in just
a
> > couple of years time will be far more ambitious than what we can
> > realistically preserve now. Is this a good reason to turn away
material
> > now?
>
> (Note that my response here is about research *data*, I have nothing
> (useful)
> to say about learning materials etc).
>
> Yes, I'm afraid it is. The reality is that preservation (beyond a
trivial
> period of time) relies heavily on a relationship between the producer
of
> the
> data and the archivist, and as time goes on, a relationship between
the
> consumer of the data and the archivist.
With regard to research data, yes I agree completely that without
metadata this material will become a problem as soon as the producer is
no longer in the picture.
However, there is an argument that we may as well keep the bitstreams
with little or no metadata on a "just in case" basis. Even if only a
small percentage are returned to by the producer (who can make sense of
them), it may well be worthwhile. Assuming of course that the cost to do
basic bitstream preservation is low in most cases. Perhaps a big
assumption...
An example I'm familiar with is at The University of Leeds which has
maintained a long term file store (effectively doing no more than basic
bitstream preservation) for many years. While I'm sure much of the data
in it has been written once and never read, there are some compelling
examples of deposited research data that has been retrospectively
discovered to be really quite valuable and recovery/re-use has been
possible as the producers have still been around.
Taking this approach further, there may even be cases where the software
of the future may provide us with better mechanisms for using the data.
Again, there are certainly examples of old data that can now be explored
and used more effectively using modern search and view software. This
suggests that the same may be true of the material we produce today.
Of course if this is the approach taken, the repository has to be very
clear about what it means to take material with a bitstream preservation
only approach!
Is this at all realistic for you Bryan? Very interested to hear your
thoughts.
>
> Experience tells us: if you don't get the producer actively involved,
> producing adequate discipline specific metadata, right at the
beginning,
> and
> if you don't then actively migrate that metadata ... you'll end up
with
> bits
> and bytes that are simply impracticable to deal with because you need
> humans
> to help ... and it's simply become unscalable. We're already in that
> position
> with some of our early datasets ... it's not that we can't read the
> format,
> it's that the information encoded in the format isn't good enough, and
we
> need a human to deal with it ... but we've got 35,000 files in that
> format,
> each could take 1-15 minutes or so to deal with. You do the maths ...
> that's
> just one format!
Can you explain what you mean by the information not being "good enough"
as opposed to not being able to read the format? Would you say this is a
bigger challenge than the format obsolescence problem?
>
> So, by saying to the producer: "don't worry, just biff me the data,
we'll
> work
> out how to preserve it later" is giving them license to think you've
done
> the
> preservation work, but you haven't, and you're going to have to come
back
> to
> it ... and by then there might be no one willing to pay for the work!
>
> (For the record: definition of "early stage": the BADC has been
preserving
> digital data for 11 years, and it grew out of a preceeding entity ...
> arguably we represent several decades of experience doing this ... and
> we've
> made a lot of mistakes ... some of which I see being repeated).
>
> > >Then they need to hold a hard line against function creep, and only
> >
> > accept >material in "well known formats" (whatever that means) with
> > "well
> >
> > This is a popular strategy to ease the preservation problems the
> > repositories are taking on. However, the reality is that when faced
with
> > a limited number of submission formats, submitters either don't
bother
> > or they perform migrations themselves. Do we really want to place
> > complex preservation actions in the hands of the users? No records
are
> > kept of what action they take and migrations are performed in an ad
hoc
> > way. This could well be creating an even bigger preservation
challenge
> > for us in the future.
>
> The bottom line is the folk you call users are the data producers, and
> they
> know more about what they're doing than we do. In particular, if
anyone
> has
> to make ad hoc decisions, rather them than me! When I can do it
properly,
> then I'll get involved.
>
> (Remember, I'm talking about research data, the arguments Paul makes
are
> quite
> tenable with other types of "preservation entity", and i'm only making
> these
> arguments to try and keep IR's - with a sensible definition -
practical
> and
> useful).
With non-research data, one of my major concerns is where the producer
takes some kind of preservation action that is not recorded. Is this
considered to be a problem with the research data that you encounter, or
is this not really a concern in the context of the bigger challenges the
material poses?
Cheers
Paul Wheatley
--------
Digital Preservation Manager
eIS/Architecture
The British Library
01937 546254
[log in to unmask]
**************************************************************************
Experience the British Library online at www.bl.uk
Help the British Library conserve the world's knowledge. Adopt a Book. www.bl.uk/adoptabook
The Library's St Pancras site is WiFI - enabled
**************************************************************************
The information contained in this e-mail is confidential and may be legally privileged. It is intended for the addressee(s) only. If you are not the intended recipient, please delete this e-mail and notify the [log in to unmask] : The contents of this e-mail must not be disclosed or copied without the sender's consent.
The statements and opinions expressed in this message are those of the author and do not necessarily reflect those of the British Library. The British Library does not take any responsibility for the views of the author.
**************************************************************************
|