Future of Academic Biocomputing Facilities

Duncan Rouch ROUCHDA at VAX1.COMPUTER-CENTRE.BIRMINGHAM.AC.UK
Tue May 4 08:02:00 EST 1993


Future of Academic Biocomputing Facilities.


Hi Netters,
If your work is touched by Biocomputing this discussion is for you.  
If you are a molecular biologist user, a programmer, or keep a 
biocomputing service on the road you have a stake in the future of 
biocomputing facilities.  Effective service requires effective 
communication between a service and its customer molecular 
biologists, in both directions.  Users can provide both necessary 
feedback and political support for biocomputing services [1].  
Software developments must be integrated into the service environment 
to be effective.  Here we discuss the future of academic biocomputing 
facilities, with emphasis on commercialization.


Duncan Rouch, University of Birmingham UK (D.A.Rouch at BHAM.AC.UK)
Tim Littlejohn, Universite de Montreal (Little at ere.umontreal.ca)



Contents:

1. Introduction: Optimizing Biocomputing With or Without Commercialization.

2. Academic Computing Facilities: Obtaining and Maintaining Success.
	2.1  Obtaining a Better Academic Service.
	2.2  Factors Against Optimization.
		2.2a  Suspended Animation.
		2.2b  Becoming Invisible.

3. Commercialization.
	3.1  Which facilities are we talking about?
		3.1.1  National Services.
		3.1.2  Site-wide Services.
			3.1.2a  Why Site-wide Services are Important.
			3.1.2b  Site-wide Services as Targets for 
				Commercialization.
4. Going Commercial.
	4.1  Going Commercial, Badly.
		4.1.1  Site-wide Services Seen as Expendable.
		4.1.2  Lack of Effective Planning.
		4.1.3  Inadequate Management Structure.
		4.1.4  A Poor Service as the Basis.
	4.2  Going Commercial, Improving the Odds.
		4.2.1  Commercialization and Public Responsibility.
		4.2.2  Starting With a Good Service.
		4.2.3  Charging For Services.
	4.3  Maintaining a Commercialized Service.
	4.4  Competition.

5. Conclusion.

6. Acknowledgements.

7. References.



1. Introduction:  Optimizing Biocomputing With or Without 
Commercialization.
 
Although some academic computing facilities have already been 
commercialized, for example what is now Minnesota Supercomputer 
Center Inc., at the University of Minnesota, this is not yet a 
widespread phenomenon.  However, at least in the UK, there are 
government moves to marketize the academic world, which could well 
lead to pressure to turn academic computing services into commercial 
enterprises.

It is essential for scientists and computing staff to discuss what 
arrangement of computing services would create the optimal support 
for research in the biological sciences, which could include an element 
of commercialization.  This is critical if we are to shape the future
of biocomputing facilities and not leave their fate to bureaucrats 
and legislators who may not fully appreciate their importance.


2. Academic Computing Facilities: Obtaining and Maintaining Success.

Our objective here is to look at how a successful computing service 
can be established and maintained, whether or not it has a commercial
component.  We define a successful service as one that performs 
effectively in five areas [1]. It: 
     (1) trains staff to deal with new technology; 
     (2) trains staff in care of user scientists (customer care); 
     (3) trains customers to use the facilities offered;
     (4) obtains and acts on feedback from its customers; 
     (5) plans for future needs.

Users of a particular facility may be research scientists, or 
science educators, or more usually of both kinds. 

Successful biocomputing facilities effectively support users in four 
major classes of computer-aided techniques:
     (a) functional-sequence analysis; 
     (b) evolutionary-sequence analysis; 
     (c) molecular modelling;
     (d) analysis of related information databases, such as literature.  

Such facilities will also benefit from sharing information and expertise 
with other facilities [1].  Also, users of small department- and 
lab-specific facilities can be helped where support for these 
facilities can be offered by a local site-wide service, which is
able to do this through economies of scale.


   2.1  Obtaining a Better Academic Service.

For a facility to provide an optimal service for a given level 
of resources it must be highly responsive to the needs of user 
scientists.  This can be encouraged, for example, in the case of 
a site-wide service where the budget for the research and education 
departments is drawn on to fund the computing service.  Then the 
departments should have some say in what kind of service they 
should get.  This gives each department a financial incentive for 
pressing for an efficient facility: a more efficient facility would
also help improve the research and educational performance of the 
department.  

A prerequisite for creating and maintaining an efficient service is 
effective management [2].  Effective management personnel: 

	(a) identify what facilities are required; 

	(b) decide what level of service can be offered with a 
	    given amount of resources;

	(c) have the authority to argue for sufficient resources
	    to offer an adequate service, if these do not already exist; 

	(d) set up services and keep them running 
	    to meet set objectives;

	(e) advance user-facility and staff-management understanding,
            by communication, through: (i) elucidation of user needs, 
            (ii) giving users practical expectations of what facilities 
            can be provided, (iii) ensuring that staff perceive that 
            managers have realistic expectations of their performance;   

	(f) keep the service up to date in relation to the changing
	    needs of user scientists.

A not uncommon failing is not to give the head manager of a service
sufficient status within the institution. A head manager must have a 
relative status that is reasonably close to that of the senior decision
makers, in order to be given due weight when arguing for appropriate
resources.  To put this another way, managers must have adequate power 
to fulfil the responsibilities of the job. This is necessary to allow 
them to make the necessary decisions to make a success of the service.

An efficient computing service among other things should have the 
capability to both upgrade facilities and user-support as new 
technology evolves, and expand the facilities to meet the increased 
need for access to electronic information [1].  Now we look at 
factors that might hinder this capability.


   2.2  Factors Against Optimization.

    2.2a  Suspended Animation.

Obviously if a computing service is suffering a restricted budget it 
is difficult to introduce change, for example, a major hardware upgrade.  
If this occurs the service may go into a state of suspended animation. 
However, if the restrictive budget is in place for a reasonably short 
period, especially following years of an adequate budget, no long term 
damage is likely to result.

On the other hand, if a restrictive budget has been in place for 
years a computer service may be in danger of failing to provide a 
level of service required for the research it supports to be 
internationally competitive.  This may eventually result in the 
future of the host institution being threatened.


An unduly restrictive budget over time may be the result of an 
inefficient management structure.  Typically, too many vertical 
levels of administration can cause an organization to move very 
slowly.  Administrations can be conditioned for a relatively 
constant environment and cope badly with the need for change.  This 
has also been found to be a problem in some large commercial 
corporations.  

The best long-term solution to a badly setup administration is 
naturally a total reorganization of the administrative structure.  
This may be a job best performed by outside commercial management 
consultants.

However, a catch with radical restructuring is that unforseen 
problems can easily arise so it can be some time before optimal 
performance of the administration is reached.  In this case it might 
be useful to target an under-resourced computing facility for extra 
aid early on in the restructuring process.


    2.2b  Becoming Invisible.

A factor that may deleteriously affect the quality of even the best
computing service is naturally a budget cut.  A small cut may be 
allow an adequate level of service to be maintained, at least in 
the short term.  

However, a major budget cut can put a service in danger of becoming 
totally ineffective.  Essential maintenance and hardware/software 
upgrading may be out of the question. Also, front-line user support 
might have to be cut entirely.  Interaction with customer scientists 
is essential for maintenance of a successful service [1].  So, without 
the interaction provided by front-line user support, combined with 
reduced hardware/software facilities, the local service may become 
totally ineffective.

In this case individual departments or labs may attempt to meet their 
own computing needs or rely on national services.  Then the local 
service may be seen as totally unecessary and cut completely. 

In the long-term, the entire loss of a local service will tend to 
result in inefficient computing support across the site.  
Financially well endowed departments may be able to cope reasonably 
well but small, less well off departments may not. 

Naturally, one solution to this problem is not to have the major budget 
reduction in the first place.  One way to reduce the chances of 
excessive financial restriction is for user scientists and staff in 
the computing service to work together to persuade the decision-makers 
what computing resources are necessary.


3. Commercialization.

There are two logical reasons for commercializing an academic 
computing service, (i) to increase its efficiency in supporting 
academic research and education functions, and/or (ii) to generate 
income for the host institution or organization.

Given the critical importance of computing facilities in sustaining 
both biocomputing and the electronic information revolution, the 
academic arena is only bettered if the efficency of computing 
support is increased.  This then should be the major factor in 
deciding whether to commercialize or not, and if so, in what form.

A result of a compromise between reasons (i) and (ii) could be the 
forming a computing business 



More information about the Bio-soft mailing list