[UFO Chicago] what draws you here?

Neil R. Ormos ormos@enteract.com
Tue, 29 Jan 2002 17:39:28 -0600 (CST)

Larry Garfield wrote:

> Ian Bicking wrote:

>> apt-get!  No, really, I've never reinstalled Debian, or
>> had the slightest desire to do so, except when I replaced
>> the main HD.  I have 1253 packages installed, and it
>> still works fine [ . . . ] Sure, maybe 75% of Windows
>> programs can be installed without effecting system
>> stability.  Just like maybe 75% of the programs I want on
>> Linux are packaged in Debian (more, actually).  But they
>> don't tell you which Windows programs are going to slowly
>> eat your system alive.

> [ . . . ] [T]he fundamental architecture [of Linux] is
> unquestionably more resilient and more stable.  The
> problem is that the cost of that increased reliability is
> you have to know a lot more about what you're doing than
> you do with Windows.  I'm not convinced that is inherent
> in the software.  I DO, however, believe that it is the
> "fault" of the FS/OS attitude.  There is a very strong "by
> geeks, for geeks" streak in the Unix-oid community at
> large (including the GNU folks, the Open Source people,
> the hard-core old=school Unix people, etc.), which I think
> hurts Linux and Free Software in the world at large far
> more than any Microsoft ad campaign could hope to
> accomplish.

Some of the trouble can indeed be attributed to the
attitudes of some people in the Linux community, but in my
experience, that has been a minor factor.  With the
exception of IRC and some similar media (where youth,
anonymity and distance encourage people to act in ways they
never could in person and expect to retain their teeth),
most Linux people seem reasonably helpful, or at least not
hostile, to newcomers.  People looking for help with Windows
questions aren't likely to fare any better, because almost
no one really understands Windows.

IMO, the major barrier to public acceptance of "Free"/Open
Source software is that the F/OS community has really only
addressed a small fraction of the total job of developing
and releasing a software product.  The parts that the
community has addressed so far include the basic system
design, and writing the code.  This is exactly what would be
expected from a community of volunteer programmers, as it
captures the problems of interest to the community and
exploits the skills available.  In essence, though, the code
is the easy part--anyone can do it, and its value is
measured objectively and empirically by whether it is useful
and works well, rather than whether it is elegant and
aesthetically pleasing.

What has not yet been addressed satisfactorily are all those
things that seem ancillary, and yet are so essential to
widespread public appeal--items such as human factors and
man-machine interface design; appliance-user-level
documentation; training; bulletproof OS installation on new
hardware; bulletproof applications installation; seamless
interoperability with the incumbent platforms; and the like.
And no, I'm not saying that Windows actually robustly
provides all of these things, but Windows' veneer
implementations work OK for many users and the depth
increases as new versions are shipped.  There are a number
of reasons these items have not been well attended to in
most "Free"/Open Source software:

  few people who like to write code find these tasks
  fun, interesting, or rewarding;

  some of the tasks are inherently tedious,
  time-consuming, and thankless;

  few people with the requisite skills or interests
  have infiltrated the F/OS community (or as a
  corollary, the F/OS community does not offer to
  those with the requisite skills or interests the
  same motivating factors as it offers to people who
  contribute code); and

  some of the tasks require extensive training and
  experience to do well, and quality is measured

If the "ancilliary" items were addressed, Linux and its
brethren could be quite suitable for non-technical users.