From: Greg Ralston <gregr@biochem.usyd.edu.au>
  To  : RASMB@bbri.eri.harvard.edu
  Date: Tue, 23 May 1995 10:19:17 +1000

Databases

Jack raises some interesting points that have relevance way outside RASMB. If an
error is published in a journal article, it is almost impossible to fix. Sure, 
you can write a correction, but who reads them? For years afterwards the error 
may be perpetuated.

With electronic data, changes may be made without anybody knowing. Errors can be
corrected surreptitiously, or openly; in either case, it is not necessarily easy
to know whether the data you have just downloaded and painstakingly worked with 
is the most recent, and therefore most likely to be correct.

My impression is that most of us envisaged this proposed RASMB data base as 
being something fairly informal. People could make what use they wanted of it; 
caveat emptor. Although I think that is what I would prefer, it is asking a lot 
to have someone maintain a database of many gigabytes of unknown (and possibly 
dubious) validity. Then what happens if somebody does start to use the data in a
way that involves a flux of money? The problems that Jack allude to start to 
crop up. It becomes more work for somebody.


Greg Ralston