🚀 go-pugleaf

rsw@cs.brown.EDU (Bob Weiner) writes:
>dc@stars.gsfc.nasa.gov (Dave Cottingham) writes:
>
>> I have unsuccessfully been trying to figure out exactly how this business
>> of supporting free software is supposed to work out.  But since there
>> are functioning businesses out there that do it, there is clearly an
>> answer.  The conventional system operates sort of like an unregulated
>> utility -- the people who own the code set themselved up as the only
>> people who maintain it, and they spread the costs of their activity
>> evenly over everyone who wants fixes.

>The present industry may have confused you as to what service should
>be.  If you think of a quality product, the producer of such product
>should feel an obligation to remove defects and prevent them from
>disrupting his customers' work.  Service should not be so much about
>defect removal but about extending functionality to meet customer
>demands and supporting them in their use of the product through
>training and consulting.  If a customer has his specific needs met in
>an economical fashion, it is doubtful he will care whether he paid
>exactly his fair share of the service company's yearly revenue.

I disagree, for the moment, on the value of getting functionality
extended and or customization added.  In some markets, the goal
appears to be consistent, uniform behavior.

Some machine at Stanford has some default printer setting that is
different from what we use at home, so that documents printed there
are not WYSIWYG.

My brother administers some number of machines at a university; it is
his goal to get all those machines looking exactly alike, so that no
(admittedly naive, but most of them are) user is surprised by
diferences in machine-to-machine configuration.

In my use of emacs (both Gosling and Gnu), my major use of the
editor's configurability is to make all the same keys do exactly the
same things, no matter what terminal I use.

In system administration, one goal of large shops is to provide better
service at lower overhead.  One way to achieve this is to make every
machine out there look as much alike as possible.

My grandfather chooses which computer and software to buy based on
what his friends own, so that he can easily get help both with the
hardware and the software.  Variation in the software does not help
him.

In the X11 world, a typical desktop has server and window manager in
different processes.  This provides great flexibility, at the expense
of performance.  Too bad I never use the flexibility.

These some examples of where custom service is not value-added.  I
don't know if these are the common cases, but I suspect that they are.
The FSF arguments regarding look-and-feel cut both ways -- what good
is a customized typewriter keyboard?

There is one case where a world of customized free software is
definitely an advantage -- if all the software is slightly different,
things like the  Morris worm/virus have a much harder time spreading.

>Think of the bigger picture and it may appear to you.  Free software
>opens up the service market and allows competitive forces to deliver
>more quality service for software products.

In the case of my car, bicycle, stereo, toaster, refrigerator, and
furnace, quality service is the Maytag repair man (i.e., no service).
I'v received "quality service" at Nordstrom, and it was nice, but not
particularly what I want to pay for on a regular basis.  When I get my
car repaired, I do go someplace where I like the service, but I regard
it as a defect in the car that it needs service at all.

Presumably, the bigger picture has appeared to you.  Can you provide
examples of where "quality service for software products" is an
interesting (i.e., profit-making) concept?

David Chase
Sun