🚀 go-pugleaf

In article <1991Apr4.162300.14422@visix.com> brett@visix.com writes:
>In article <1991Mar28.215825.18142@federal.uucp>, hecker@federal.uucp
(Frank Hecker) writes:
>> In the world of large-system software, extensive service after
>> installation is required, and for legitimate reasons.
>
>True, but what I find is that "large-system software" installation procedures
>do not take in account many issues, and are written for a very small number of
>machines/hardware configurations.  IMHO this is the fault of the programmers
>and should NOT have to be "made-up for" by the customer paying all sorts of
>money for on-site installation and configuration help.

I should have made it more clear that I was thinking not just of the
typical UN*X site with workstations, NFS servers, etc., but also of
large mainframe and minicomputer sites with lots of oddball systems
and connections to be made between them.  (For example, in a prior job
I had to customize an RJE interface that used UNIVAC 1004 protocol.)
Many of these type of issues are unique to a particular site.  (Of
course, open systems and standards are supposed to solve all this :-)

>> I believe that the business of providing service for freely available
>> software is justified and has potential to the extent that
>> organizations grow to depend on the software for core business
>                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>
>Well, I don't think that any business should depend on free software for
>core business.  You are just asking for problems.

Why?  The trust you put in software is a function of the quality of
the original code and the time you spend ensuring that it works in
your environment.  One could argue that the first is only loosely
correlated with the original source of the software, and the second
not at all.
-- 
Frank Hecker
...!uunet!tdmfed!hecker