Re: Biology packages in FreeBSD

[Bill Pearson <wrp@unix.mail.virginia.edu>]

Thank you for including me in your discussions on Debian distribution
of bioinformatics programs.

I have been distributing Bioinformatics programs for almost 30 years
(I distributed some programs back in the mid-70's on punch cards), and
the FASTA program (and its predecessor, FASTP) has been distributed
for about 20 years.

It has mostly been my experience that the difficulties that biologists
have using FASTA, or programs like them, would be reduced very
slightly by providing a binary distribution.  It is very easy to
compile FASTA - the only libraries required are the standard 'C'
library, and the math library.  So different architectures, and
different versions of Linux, really don't pose a problem.

It can be difficult to install FASTA in a way that makes it easy for
users to find the sequence databases.  Not only do the databases need
to be downloaded, but users must understand something about the 
different sequence flat file formats, and, for things to be really
easy, a configuration file must be edited to list the names and
locations of the sequence databases.  Most people who have problems
with FASTA have problems at this step.

But FASTA sometimes has bugs (which are usually rather subtle), at
which point it is helpful for users to contact me.  Genuine bugs are
of great concern to me; I try to fix them within a week.  My
experience with FASTA WWW sites is that users sometimes find a
problem, but do not tell me about it.  I like people to get the
program from me, so they can contact me about it.

Often, the first thing I will do is suggest that they download a
current version of the program.  That's a sensible suggestion if they
got the program from me originally, but can be quite confusing if it
came as part of a Linux distribution (does updating their Linux
distribution update FASTA?).

Thus, I believe that scripts that download the current version of
FASTA, and databases (databases are available from ftp.ebi.ac.uk/ and
ftp.ncbi.nih.gov/) are more useful for most Bioinformatics users than
FASTA binaries.

Bill Pearson