On 7 Sep 2000, Bradley K. Sherman wrote:
> In article <Pine.LNX.4.10.10009040852150.12440-100000 at riker.ebi.ac.uk>,
> Ewan Birney <birney at ebi.ac.uk> wrote:
> >
> >Where should ACEDB go?
> >
>> More documentation, rich in examples. Tutorials at
> major conferences.
>> ACeDB is still the only system where middle-aged,
> tenured Molecular Biologists sit down and understand
> the interaction quickly.
I've been an acedb database curator now for almost a year (managing an
Arabidopsis genome database as part of a network of crop plant databases).
Part of my job is also to run training sessions for existing and
potential users of these databases.
It seems to me that very few end users will bother/have bothered to
download databases to install locally. Many people might only want to
occasionally use one of these databases, invariably to find out more about
their sequence of interest (particularly in a genomic context). I'm not
sure how much the windows port of acedb will encourage people to install
it locally, but the problems remain that some of these databases are
getting very big now.
My database (AGR) has reached about 2.5 Gb in total (partly caused by the
problem (bug?) that acedb 'puts on weight' when you keep on updating
entries, even if you remove entries! - I've found that dumping out the
entire database and reloading can shed up to 0.5 Gb). I guess the C.
elegans database is reaching a similar size. I can't envisage many people
setting up these databases if they have to first download so much data.
Therefore it seems that a web based approach will become increasingly more
popular, but then you run into the limitations of AcePerl, AceBrowser, and
WebAce - particularly when interacting with images from gifaceserver.
I think having a decent manual (downloadable as a pdf?) would help a lot,
the current collection of online manuals and guides are often woefully out
of date, and a centrally maintained, regularly updated, manual at the
Sanger Centre would be a useful resource to point potential and existing
users to.
If acedb wants to remain a popular and practical choice for managing
biological, and particularly genomic data then it must (at the least) be
able to handle the torrent of data that is starting to be produced by
various post-genomics projects (esp. microarray data). As these data,
along with the ever increasing amounts of sequence data, mean that acedb
databases will just get bigger and bigger, then it seems to me that it
will become harder to expect people to download and install databases
locally...I guess this all hinges on whether growth of genomic data
exceeds growth in bandwidth capacity/disk storage.
Just my two-penneth-worth.
Keith
~ Keith Bradnam - Developer, Arabidopsis Genome Resource (AGR)
~ Nottingham Arabidopsis Stock Centre - http://nasc.nott.ac.uk/
~ University Park, University of Nottingham, NG7 2RD, UK
~ Tel: (0115) 951 3091
