These were some of Reinhard Doelz's recent comments:
> .
> .
> .
> Don't understand me wrong but I consider a comprehensive package
> to be a software environment which is (a) smoothly integrated
> (b) well supported and (c) widely used so that there's a considerable
> experience, and the chance that you do not need to educate incoming
> newbies. Fuzzy users and researchers are not necessarily the same
> community - why should a molecular biologist know how to configure
> a software package if the hardware spec is not quite the same as requested
> by a program package, or why should a molecular biologist write a
> small program to transfer data from format X to format Y ?
> .
> .
> .
> Keep these restrictions in mind - public domain software is great as
> long as you are not dependent on it. Use it as much as you can if you
> feel that you benefit from it, but do not expect that you trash
> a XXXX$ license and get replacements for free at 0$ cost. It might
> be 0$ cost on the purchase but much higher if reevaluated with time
> of support staff, research staff, and time lost due to incomprehensive
> results obtained due to unpractical reformatting procedures.
> .
> .
> .
For what it's worth, I am one who totally agrees with Reinhard!
It's very easy to focus on certain, written-down fixed costs (e.g.,
the price of a GCG license) while ignoring more intangible but much
higher variable costs -- like the cost of manpower, which by far
exceeds the cost of almost everything tangible.
I have no experience with the IntelliGenetics package, but my
experience with the GCG package -- the consistency of the programs,
the clarity & comprehensiveness of the documentation, and most
notably the efficiency & accuracy of their technical support -- has
been GREAT. All of those factors have certainly saved me and my
hundreds of users 1000's of hours of labor which would otherwise
have been spent figuring out how to run diverse non-commerical
programs, how to modify their executions, how to properly prepare
input and interpret output, how to "transfer data from format X to
format Y", etc.
Unless there is a very tight group of scientists with a precisely
defined need for a small subset of sequence analysis functions, it
is virtually certain that trying to save money by relying on a
collection of multiple free programs will cost a lot more money --
expressed in many people's lost labor hours -- in the long run.
In all information-management work, the largest single fraction of
time is spent on "miscellaneous" tasks. The object is to keep that
fraction at a minimum.
Mark Reboul
Columbia-Presbyterian Cancer Center Computing Facility
mark at cuccfa.ccc.columbia.edu
