I have been following the discussion of the relative merits of various
media for archiving image data. Perhaps this is not part of the original
question, but no one has addressed a more serious problem with volume data
storage: how do you go back and find something that you did three years
ago? What about twenty years ago, as Paul Goodwin suggests? We may have a
rough idea where to look in our own data record for a particular result,
but what if we have to find the results generated by someone else who is no
longer with the lab/company? How many of you annotate those CDs you burn
with some readable record of what the contents represent?

If we ever expect to retrieve any random piece of our data, then we need
some form of searchable database of our results. But what is the unit, the
"record" of this database? An image? A set of images, such as a ratio time
series or a confocal stack? An experiment? A folder of images and graphs
extracted from them? Do we database everything, or do we apply some
selection criteria to keep the database to a reasonable size?

If we need to retrieve something, most of us are forced to fall back on
that most primitive of our scientific tools - our laboratory notebook. This
ancient tool evolved in a time when there was no other method for recording
data than putting pen to paper. This may be scientific blasphemy, but I
think that the laboratory notebook is becoming more of a hindrance than a
help in research which for many of us is performed entirely on scientific
instruments and whose data is exclusively processed on computers. Does no
one else see the irony of having a lab notebook filled with pasted-in data
spreadsheets and images generated by computers, where about the only thing
handwritten is the date and signature? I mean, wouldn't it be easier to
type in the few non-computer-generated data items into a running computer
record than to do what we do now, which is print it all out and paste it in
to our notebooks? And please, don't tell me about the legal reasons for
keeping a notebook - I already know them, I work in industry. It is just
that I am now contemplating the need to keep two sets of records - the
(legal) static paper one, and the (useful) dynamic, searchable, updateable
computer one.

Thanks to developments in the software industry, we are now very close to
having a decent method for keeping an "on-line notebook" - one with not
just text, but charts, images, sounds, links, moving pictures, you name it.
Of course, this notebook will be searchable. Few of us had even heard of a
word processor or a computer graphics program twenty years ago, yet we now
find them indispensable. Perhaps we will undergo a similar transition in
the next twenty years with regard to how we maintain our laboratory
records. To bring the discussion back home to archiving of data, I'll just
say that it seems obvious to me that the record of our laboratory work and
the catalog of our stored raw data are intrinsically linked, and that since
the stored raw data is not on paper, the other components shouldn't be
either.

I would enjoy hearing other people's ideas on these topics, or pointers to
related material. I don't want to start a major thread that is unrelated to
the topic of confocal microscopy, so please reply directly to me if you
want to discuss this. I will gladly post a synopsis of any discussion to
this group.

Thanks for the bandwidth,

Kurt

________________________________________________________________________
Kurt M. Scudder, Research Scientist          tel: +45 4442 1412
Novo Nordisk BioImage                             fax: +45 4442 1411
Moerkhoej Bygade 28                               email: [log in to unmask]
DK-2860 Soeborg, Denmark