Path: utzoo!utgpu!watserv1!watmath!uunet!wuarchive!uwm.edu!bionet!FCRFV1.NCIFCRF.GOV!gribskov
From: gribskov@FCRFV1.NCIFCRF.GOV ("Gribskov, Michael")
Newsgroups: bionet.molbio.bio-matrix
Subject: real science, science, genomes etc.
Message-ID: <9102111606.AA25622@genbank.bio.net>
Date: 11 Feb 91 16:03:00 GMT
Sender: daemon@genbank.bio.net
Lines: 73

In response to the recent vitriolic and contradictory comments by
Ellington@frodo.mgh.harvard.edu (Deaddog) 

>Yes, this is the crux of the matter.  Is "real science" more important
>than bookkeeping efforts such as the Genome Boondoggle? > >"Real
science" generally addresses questions (the old "test the >hypothesis"
saw). 

It seems to me that Dr. Ellington expresses a rather non-mainstream view
of what real science is.  A look at some of the more mainstream
philosophy on this point, such as Kuhn or Popper, might be illuminating.
My feeling is that "Real science" involves quite a lot of bookkeeping,
checking on the accuracy of models, and that they perform according to
predictions.  If you take a more Baconian view, sequencing the human
genome is the ultimate in real science and we should hold off on making
inferences until we have completed a substantial part of the sequence. 

I suppose the that the cataloging of galaxies is a similar boondoggle,
in spite of the fact that this effort is currently leading to some of
the most important and interesting progress in astrophysics.  I guess
the real problem with these kinds of projects is that the day-to-day
work is tedious, and results only come in the long term.  Strange how
much of science falls in that category isn't it? 

>I want to do "real science" on questions that I think are interesting. 

>Most researchers put up with anonymity and low pay and the constant
>hassles of teaching and grant writing so that they can Do What They
>Want To Do.  And to suggest otherwise is ridiculous and, in fact,
>insulting. 

These comments make it clear that what Dr. Ellington considers to be
real science are only problems that are "what I think are interesting".
Anything else is clearly a boondoggle. 

I suggest that in fact researchers are not that poorly paid.  Most
faculty appointments carry a salary that is substantially above the
median salary for the US.  Even post-doctoral salaries put scientists in
the middle class.  In fact, it is primarily in comparison to salaries in
industry that scientists are poorly paid.  This seems somewhat unfair
since "real science" is apparently not done in industry. It is true that
many scientists put in long hours, but this is mainly for fame (albeit a
limited kind).  How many people who stay on in science if papers could
only be published anonymously? 

Modern science is too complicated for every person to be out there on
their own.  We need to have people interested in methods development,
and in assembling large amounts of data into useful forms.  Some people
have suggested that in the near future there will be a critical need for
people who do no "hands on" research, but mainly concentrate on
synthesizing results from specialized researches who are two narrowly
focused to see the forest for the trees.  In the current organization,
any of these important activities are only fundable only if they are
tacked on to a "research" proposal tackling "real questions".  But let's
be realistic -- are grant proposals generally a statement of a real
problem -- or are they a post hoc justification for doing what we think
is interesting. 

Let me just mention that I spend most of my time doing what Dr.
Ellington might (or might not) consider to be real science (x-ray
crystallography).  I have no association with any genome center, nor do
I get funding from any genome initiative related agency.  I do find that
there is a lot of useful information to be extracted from the currently
available sequence information. And although I have my own misgivings
about the genome project, and especially about the rate that funding is
being ramped up, overall I think it will be an extremely useful project.
 One of the main positive features of the genome project is that for
once it is explicit that the people developing methods, and those
putting the data in useful forms are valuable members of the project. 
About time! 

Michael Gribskov
gribskov@ncifcrf.gov