Support for Database Maintenance

[James H. Beach]

Norm Penny wrote:

>                 Because of institutional budget constraints, any major
>           collection is faced with either not doing collection
>           databasing, or severely curtailing most other scientific
>           support activities.  I still think that the best solution is
>           for NSF to support maintenance of databases.


[With apologies to the non-US subscribers who tolerate these US
funding discussions]


NSF is always stretched to the budget limit for infrastructure costs.  I
predict that NSF will never be able to support more than a small
fraction of the ongoing maintenance costs of large, heavily-used, multiple
institution, community databases -- let alone the system maintenance,
data management, editing and "curation" associated with isolated
institutional systems.

We have an infrastructural program that is maxed out every year with
maintenance of existing long-term projects, with essentially no funding
available for new efforts.  That is what happens when a community depends
on flat NSF budgets to cover ever-increasing maintenance costs, be it for
living stock centers, databases, observatories, SEM facilities or
whatever. (But if you are arguing that NSF's budget needs to be increased,
I'm with you!)

Speaking from the perspective of my infrastructural program, Database
Activities in Biology, we are very concerned about this issue and are
working to find creative solutions to it, but there is no easy answer.
Economies of scale are one avenue. The US biological museum community
cannot afford to support a unique database management system in each
department at each institution.  Nor are we likely to all interact with
single, centralized, monolithic system.  Database architecture options
will be strongly tempered by cost considerations and are likely to fall
between those two extremes.  Fortunately the networking and information
technology communities are driving toward more cost-effective ways of
supporting distributed computing and information retrieval across multiple
institutional systems.  Access tools are going in a similar direction,
like with this thing called "Netscape."

There are many financial and infrastructural issues to be resolved with
collections community database efforts, the high cost of software
development and maintenance is a good one. Also it seems inevitable that
professional societies, that have long subsidized print-based
communications for their membership, should subsidize appropriate online
services, even databases, for the use of their members.  The discussion
thread about online databases replacing print-based species descriptions
is interesting for several reasons, one being that it would scale nicely
from an economic perspective.  Consider the departmental, museum and
professional society subsidies now going into publishing and distributing
new species descriptions on pulp-based media. (But I am straying from
your original topic.)

In my opinion, the worse thing the collections community can do for itself
is to consider NSF awards for new network and database projects to be
"Trojan Horses." Are three-year research awards "Trojan Horses?"

Finding ways to support ongoing costs is a problem but network access to
well-curated collection and classification information is an essential
survival strategy for the discipline and its collections.


______________________________________________________________________
Dr. James H. Beach                                 Tel:   703 306-1470
Program Director                                   Fax:   703 306-0356
Database Activities, Room 615                   E-mail: jbeach at nsf.gov
Biological Instrumentation and Resources
National Science Foundation                           E-mail preferred
4201 Wilson Boulevard
Arlington, VA 22230