Subject: Re: Kent Beck's talk
From: Adam Turoff <ziggy@panix.com>
Date: Wed, 4 Aug 2004 23:42:24 -0400

On Wed, Aug 04, 2004 at 07:11:56PM -0400, Ian Lance Taylor wrote:
> Adam Turoff <ziggy@panix.com> writes:
> > Fifteen years ago, if you built a better lint, programmers would be
> > happy to pay an hour's wages for the tool, enough so that you could make
> > a living and work on it full time for years to come.  
> 
> That does not accord with my recollection of life fifteen years ago.

A couple of people have taken issue with my casual assertion.  Here is
what I was talking about.

About 15 years ago, I remember the programmer mags were full of ads for
development *tools* (not just libraries, but full fledged tools like
lint, C compilers, and so on).  I remember one job where we were
developing on DOS and had to buy a third party linker just to get our
code to load into memory.  I also remember Gimpel Software's ads for
PC-Lint.

Back then, if you were doing any real development, you purchased a
compiler and possibly some auxiliary tools and/or libraries.  That was
the norm for a good deal of the world.  Buying tools like revision
control, database servers, lint, text editors (with syntax coloring!) 
and the like were quite common.  

Eventually, the add-on tools got bundled into the core development
environment to drive upgrades, switching vendors or even switching
languages.  And the open source tools got better and more integrated as
well.  And the price on programming tools dropped to the floor to the
point as a general rule, they are all available free-of-charge, except
in some edge cases and specialized markets.  Even Microsoft's core tools
are free-of-charge today.  

In the process, a tool like JUnit that could have easily fetched, say,
$100/seat fifteen years ago, is open source today.  And there is an
expectation that anyone would use a tool like JUnit (or Ant) today
wouldn't pay for it if it were for sale.  They'd suffer without, use a
different open source tool, or start writing an open source tool to
replace it.  


My point remains that at one point in time, Kent Beck *could* have made
a living working on a tool like JUnit.  From what I remember about 1989,
if he had released a product to help programmers write better software
(on the order of ~$1B/year in productivity gains), he would have been
able to have his "whole job" focusing on JUnit, earning enough money
to pay the mortgage and his daughter's tuition.  

But it's not 1989 anymore, and something is different, to the point
where even if he is truly delivering $1B in value to the industry as a
whole, there is no expectation that working on a tool like JUnit should
provide him any income whatsoever.

Z.