[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]


	 Date: Thu, 17 Aug 89 11:02 EDT
	 From: barmar@Think.COM (Barry Margolin)
	     Date: Thu, 17 Aug 89 08:53:08 CDT
	     From: mac%cadillac.cad.mcc.com@mcc.com (Mac Michaels)
	     With a policy like this it kind of makes me glad that I no longer use a
	     Symbolics.  Such a policy is warranted only when you ship software and
	     documentation with absolutely no bugs, errors, or inconsistencies.  It
	     seems only fair that you pay people who report bugs and their solutions a
	     royality each time you sell that information.

I assume that Symbolics accepts bug reports sent in by people without
software support.  By accepting I mean that they put bug reports on a
list of things to be fixed eventually.  Is this the case?

	 That's a highly idealistic attitude, and businesses work on profit, not

Nobody asks for altruism.  But when I buy a software product, I pay a
certain price for it.  When it turns out that the product I got for my
money doesn't work and I have to pay additional money before I get a
working product, then this doesn't seem fair to me.

Also don't forget the cost of bugs to the customer.  A bug in the
software you use can cost you a lot of time, especially if you're not
wizardly enough to immediately tell a bug from a feature you may not
understand.  How long do you try to make sense of something until you
decide it must be a bug?  Especially if you don't have access to
source code?

		    Does your current vendor (Sun, I presume) pay you for your
	 bug reports?

	 First of all, not all calls to SW Support are to report bugs; many
	 times, it is customers asking for assistance.  Why should they get such
	 help without paying for it?  Sometimes the customer is looking for
	 patches that enhance the system for the purposes of specific

	 Second, why should software be any different from hardware?  

     Software is *very* different from hardware.

     Hardware is normally tested before it is shipped.  

Software is also tested before it is shipped.  The difference is in
the complexities.  Testing a large proportion of all possible
situations seems to be easier when you test hardware.  Testing
software is a colossal task.  We all know that.  Software companies
use elaborate strategies to achieve a certain level of correctness:
e.g., software is tested by the software developer, shipped out to
customers who serve as alpha and beta test sites, etc..  However,
these methods are often not sufficient, and software products require
a long ``life'' (many years) before they are virtually bug-free.  The
7.0 version of the Symbolics presentation substrate, for example, was
full of bugs.  The 7.2 version is usable for most purposes.  The 7.5
version should be even more mature.  (I can't wait to see if the bugs I
reported will be fixed.)

							If it fails in the
     field (after some warantee period) then you should pay to have it

     Software bugs are flaws in the original product design and they are the
     result of incomplete testing.  Bugs are shipped with the product, they
     do not "appear" with time (unless you suffer from bit-rot!)  It is quite
     possible that certain bugs are not manifested for some time, depending
     on a particular user's exercise of the software affected by the problem.
     However, discovery of a bug after the warantee period still reflects a
     flaw in the original design.

Software doesn't wear and tear in the same sense as hardware.  The
kind of aging I can think of right now is compatibility with new
software versions.  For example, old software that used Zetalisp style
characters and strings had to be adapted to CommonLisp style
representations of characters -- that is what I would call software

     The automotive industry calls a bug fix a "recall".  The automotive
     companies almost always pays for the fix, not the customer.

This exactly corresponds to my gut feeling of right and wrong.  (Am I
naive, or what?)

     I guess I am always amused at how we can let companies sucker us into
     buying a defective product and then get us to agree to pay them to fix
     the problem.  (Doesn't this sound like the present state of Pentagon
     weapons systems purchases?)

No additional comments necessary...

	 If I don't
	 pay for a service plan, and a board fails, I'd have to pay Symbolics to
	 send someone to replace the board.  

But what if the design of the board is flawed in some way that causes
it to break more often?  My Macintosh doesn't fail nearly as often as
my Symbolics.  Should I have to pay for these design flaws?  If we use
the automobile analogy, I guess, the answer is yes :-(.  If you buy an
American car its just going to be in the shop more often than a
Japanese car.

					     By the same token, if you don't pay
	 for SW Support, and you discover a failing function, you pay for
	 Symbolics to send you a fixed version.

	 Software support is a service, not a right.

Software support covers many different things:

     For assistance in using the product, I agree.  For obtaining new
     capabilities, I agree.  For fixing design flaws, I strongly disagree.  I
     have a right to expect it to be fixed for free.

I'm quite sure that Symbolics and other companies are legally entitled
to sell buggy products without any obligation to fix them.  However
that could be changed either through market forces (i.e., customers'
refusal to buy those products) or congressional regulation.  Because
of the current legal situation, companies have a low interest in fixing
bugs; getting new products out brings in the money - not fixing things
that people have already payed for.

... andreas ..:-).