Subject: Re: Software quality: free vs. proprietary
From: "Stephen J. Turnbull" <>
Date: Tue, 26 Oct 1999 02:55:35 +0900 (JST)

>>>>> "Ben" == Ben Tilly <> writes:

    Ben> I don't know what theory you are going by here, but
    Ben> pragmatically it doesn't work that way.  Compare the
    Ben> unwillingness of peopel to have anything to do with Sun's
    Ben> Community Source License versus the popularity of the GPL.

What people?  Perhaps you are ignoring the people _inside_ Sun and
their corporate partners?

Of course the GPL is popular.  Free software hackers are generally
also free software advocates.  That is not the whole story.

    Ben> Sure in some sense, you can get an arbitrarily high fraction,
    Ben> but the fraction you get is strongly dependent upon how close
    Ben> you are to what most people call "free".

If Bill Gates tell his employees that part of their job is to keep up
their internal source archives, and pays bonuses to people who submit
patches to other departments software, I don't think they will much
care what the external license looks like.

Why do you think that approach to internalizing open source advantages
is impossible?  Steve McConnell thinks that when you're counting eyes,
"many" is on the order of 4--6 (the number of members of an inspection
team, IIRC).  If so, a moderately large development organization can
do open source _in house_.  If open source methods are really that
effective, it will be worth telling the pointy-headed bosses that
"yes, your programmers ARE supposed to be moonlighting on other
departments' projects EVERY Friday ALL day, and you are STILL
responsible for meeting your deadlines and cost objectives WITH YOUR

Certainly you lose the advantages of giving source to the users.  But
if you think about it, those advantages have the highest leverage on
quality if you expose your users to bugs early and often.  Is that
really a great idea?  (There are other reasons for doing that; but we
were discussing quality.)

    Ben> Tell me what taillights were being chased by each of the
    Ben> following pieces of software:

    Ben>  - Emacs
    Ben>  - patch
    Ben>  - BIND
    Ben>  - qmail


    Ben>  - Perl
    Ben>  - Apache

(CERN & NCSA httpd)

Of course, sendmail and CERN & NCSA httpd are free software too.  And
what do you call python, slang, ruby, bash, zsh, csh, ksh, smail, jed,
microemacs, etc, etc?  I didn't say that all the taillights were

Yes, yes, yes, of course a lot of original software was produced by a
lot of very original people.  Look at the authors.  Larry Wall is in
there twice; according to Russ Nelson, Dan Bernstein produces bug-free
software (Watts Humphrey says to expect a pro to inject 50+
defects/KLOC; what does that make Dan?); I bet the authors of the
original BIND account for 10% of current RFC count among them; RMS is
RMS.  If you could get them to work for Microsoft, they'd write great
software there, too (although I'm sure you'll argue it'd get mangled
on the way to market, that is definitely contestable).  Some of them
produced great software working (indirectly) for the DOD (if I have my
history right).

We need to talk about a process that doesn't depend on a few
superstars.  We need to talk about a process that potentially scales
to 50% or more of the market.  Eventually we need to talk about a
process that produces Windows NT, except on schedule and bug-free.

What share of the 95% of software (was that your figure, or was that
somebody else?) that is written entirely for internal consumption
are Emacs, patch, BIND, qmail, Perl, and Apache?

Oh, and for the moment I will concede the basic Internet
infrastructure apps as free software.  However, it sticks in my craw
to class DARPA-funded projects with Apache, especially from the point
of view of FSB.

    Ben> "Cannot outgrow"?  Look at the ages of the software items
    Ben> that I just listed!  While free software is certainly
    Ben> spreading today into new (for it) areas, it has a long
    Ben> history that you do not appreciate!

The anecdotes are cool.  I like them, I quote them myself when it
suits my purposes.  Statistics would be nice, though.  What I base my
statement on is vast tracts of application space where there ain't no
free software yet.  Maybe your Internet/development/sysadmin anecdotes
outweigh the internal development, shrink-wrap, stats/math, office
automation, games, etc markets put together.  I don't think so at this 
point in my studies.

    Ben> Compare gcc with Visual C++.   # never used the latter
                                        # is it best of breed or just
                                        # the Microsoft entry?

    Ben> Compare Emacs with TextPad.    # not the same class of app,
                                        # unless you mean they require the
                                        # same amount of memory if you
                                        # include all the Windows routines

    Ben> Compare TCL/Tk with VB.        # I'd rather avoid learning either
                                        # as you pointed out, I'm not a
                                        # pro programmer.
                                        # [X]Emacs Lisp is what I use
                                        # for "visual programming"

    Ben> Compare (name your favorite shell) with DOS.  # 4DOS? <duck>
                                        # XEmacs? <sidestep>
                                        # ya got me this time, although
                                        # proprietary programmers using
                                        # csh can sure produce buggy
                                        # install scripts

    Ben> You were saying?

Compare SAS with (help me, Karsten, "r"?)
Compare GAMS with (nothing comparable)  # this is infinite quality superiority
Compare Powerpoint with Magicpoint
Compare Mathematica or Maple with gnuplot
Compare WinReader Pro with strokes.el
Compare Microsoft Word with wp2latex
Compare WordPerfect with word2x

(for the last four, limit comparison to the functionality available in
the free software).

We can do this anecdote stuff indefinitely.

Oh, one more.  Let's try this one again.  Same question:

Where proprietary software succeeds, free software may continue
development, but evidently in the eyes of consumers free plus bug-free
(by hypothesis) does not outweigh whatever it is that the proprietary
software is offering, presumably at a higher monetary price, besides.

So what's the relevant quality measure there?

University of Tsukuba                Tennodai 1-1-1 Tsukuba 305-8573 JAPAN
Institute of Policy and Planning Sciences       Tel/fax: +81 (298) 53-5091
What are those two straight lines for?  "Free software rules."