On Fri, May 21, 2010 at 2:28 PM, Magnus Larsson <[log in to unmask]> wrote:
> As I understand it, the way to manage a dataset is to use tags, that so to
> speak provides the functionality that relations does in a database - linking
> stuff together with other stuff. That means to me that I need to set a
> number of tags to most items, and do so in a rather systematic way.
> Arranging sets, cases, sites, persons, types of interactions observed etc.
> As well as categorizing status of files (transcribed, transcription in
> progress, not transcribed, etc.). some of this can be done through naming
> and file endings (for instance, mp3 being a sound file).
My own thoughts - based at first glance on your notes in the last
three sentences of this paragraphy - is that your requirements would
best be met using a relational database because your requirements go
beyond management in the simple sense, and into "control" in a more
complex sense. The essentially simple structure of the bibliographic
database simply wouldn't be suitable. You could (obviously) define
special records (say, "TranscriptionProgress" with valid values of
["Transcribed", "InProgress" etc] and then repeat this type of thing
for other fields. Or you might have special meanings for standard
fields (Say, "Author" which might have say three names and the reader
understands that the first name is the interviewee and the second the
interviewer and a thrid name as the subject of the interview. But
these would essentially cumbersome workarounds - which I have
personally used - and which would, in my opinion, be unsuitable for
anything but very small, temporary projects; they would certainly not
be suitable for widescale collaboration.
The other point that occurs to me is that at least part of your needs
fall into the analytical field. Consider "types of interaction
observed"; this is (I presume) based on an analysis of (say) a video
file. There could be 100 types of interaction of which ten were seen
in a particular video and so on. Two years ago I wrote and have since
been maintaining a database program for a ethologist friend who is
studying the gold mole (Chrysochloridae fam) in its natural habitat.
This project's requirements would appear to be ver roughly similar to
your own. We did try a simple bibliographic database (EndNote) in the
first instance and this worked very well for keeping track of the
files (video, audio, ground-penetrating radar etc) _per se_ but proved
inadequate for deeper - technical - study.
There's no point in going for cheapness and simplicity in the database
system if one is thereby required to a complicated, unwieldy "naming"
system. The only "easy" way I can see around this problem is to use a
system of rigidly defined notes (included in the Notes field, say); I
have used/am using such a system now for work being done with students
all over Africa but it is very definitely not something you should
consider doing unless you're forced into it as we were in my cases.
Mike
Michael Mellody
|