Subject: Re: gdb
From: (Kragen)
Date: Thu, 18 Dec 1997 12:40:36 -0500 (EST)

I will preface this post by saying that I have the greatest respect for
Mr. Tiemann's grasp of the business world and technology, and I recognize
that it is much better than mine.

Nevertheless, I think it possible that I might have something to
contribute here.

On Thu, 18 Dec 1997, Michael Tiemann wrote:
> You are making a totally different argument, one which fails for other
> reasons.  If you distribute knowledge across everybody in the
> organization, you do minimize risk, but you maximize overhead, and the
> Cygnus model wins again because that overhead _is_ our business, not a
> distraction to our business.

I believe there may be a middle way here that is better than either of the
extreme alternatives.  A few people can spend much of their time with free
tools; much of their knowledge can diffuse to others in the organization
without a great deal of effort or time on the others' part.

In fact, this is what happens anyway, in the businesses I've been in, with
or without outside consultants, and regardless of whether or not the
software involved is freed.  Some people learn a lot about the software,
sometimes they write some of it down, some people learn a bit, including
all the most essential information, often from consulting the people who
learned a lot.

> I'm not saying that Cygnus is the only company that can do what it
> does.  What I am saying is that if you want reliable expertise, it's
> better to get it from people who focus on providing that expertise, not
> people who happen to have it or people who happen to learn it.

What I generally want is a cost-effective solution to a problem.  Some
level of expertise in the problem and the tools used to solve it will
certainly help; for any particular solution, though, there's a point
beyond which additional expertise is not very useful.  So I don't think
everyone who ever uses a piece of the GNU toolchain ought to hire Cygnus