[bip] [paul.joseph.davis at gmail.com: Re: Blog post on bioinformatics and Python]

[Andrew Dalke]

> From: Paul Davis <paul.joseph.davis at gmail.com>
> Nice post. Though, my thoughts on the rise of bioperl differ somewhat.
> Anyone else have thoughts?

As I recall, Steve Chervitz and Steven Brenner were early on
involved with organizing bioperl, but my memory is vague without
doing the research.  At first it was a collection of scripts.

With the Heidelburg ISMB (and the eclipse) it became more organized.
I think because it was being used at Sanger and EBI as part of the
HGP project that there was the institutional support, ie, "part of
your job is to develop bioperl code that others use for their work",
that made bioperl more unified.

Lincoln Stein was influential in that.  He did the perl/sybase
integration for perl4 (as a fork called "sybperl" because perl
didn't have any extension support at the time).  Sybase at the
time was the big database system for sequencing work.

But don't overlook the influence of the web.  The sequencing
work was almost all public, with a pretty strong drive to share
the data.  In the 1990s *the* programming language for web
development was Perl, especially for those with Unix experience.
I like to point out that the two main modules for CGI programming
on are  cgi-lib.pl (for perl4) by Steven Brenner and CGI.pm (for
perl5) by Lincoln Stein.

So we have:
   - perl is great for strings (compared to C/C++)
   - perl is great for unix people brought up on awk/sed/...
   - perl works with Sybase
   - perl was the best for web programming
which meant that Perl was a great choice for bioinformatics
in the 1990s, and it reached critical mass.


>

				Andrew
				dalke at dalkescientific.com