Unlimited software warranties

[John G. DeArmond]

Preface:

	Lest anyone get the wrong idea, Dick and I are friends.  Therefore
	this discussion should be considered a traditional debate and not a 
	flame.  Other thoughtful input is of course welcome.


rcd at ico.isc.com (Dick Dunn) writes:

First, about satisfaction-guaranteed warranties.

>No, it's an indication that I don't think people will pay what it would
>cost to provide the sort of support you need to back up the guarantee.
>You can't just slap on the unlimited guarantee without adding support/
>maintenance effort, and somebody has to pay for that.  In order to stay in
>business, it's the customer who has to pay for it.  Suppose it cost twice
>as much?  How long have they got to recoup their investment in a 2.5+ Mach
>based product before everyone is looking for the next "new/improved" one?

>I still have worries about an unlimited-term guarantee on something that
>becomes obsolete (without wearing out) in a year or two.  But I'll defer
>that (and what I mean by obsolete) for just a bit--so hold the flames just
>a moment.


I assert just the opposite.  A properly designed product, properly documented
will not need massive support.  Let's use the automotive industry for a 
comparison.

A combination of consumer pressure, government regulations and the competative
climate have pushed the auto business into a situation quite similiar to
software and hardware companies.  To wit, the "welded hood" concept 
(The hood is conceptually welded shut, therefore no maintenance is necessary
or performed.) has pushed car quality to the point that on average, a car
will last long enough with little enough maintenance that the auto companies
must find other ways to get you to spend more money on cars.  Furthermore,
the market will no longer tolerate much of any maintenance at all.
My wife views her Toyota as a transportation appliance.  She is pissed when
she has to do any more than change the oil.  So far, she's only been 
pissed once.

Let's look at some concepts and see how they apply to computers:

	   Cars                                       Software
------------------------------------------------------------------------------
Needs regular purchases by                      Low cost upgrade policies
    repeat customers encouraged by
	trade in allowances.

Extended warranties (7 year, 70,000 mile, etc)  Free technical support.

Lemon laws and other consumer                   ????? Comming in the future?
	protection provisions.                          Forced money-back
													guarantees.

Satisfaction guarantees (Lee's promise,         Satisfaction-guaranteed    
	Japan Inc's accomplishment, for example.)       Warranties.

Consumers' desire for bells, whistles           Same
	and gadgets at low cost.

No obvious defects or design flaws              No major bugs and certainly
	(fixed at no cost if they slip in)            no software-induced system
                                                  crashes.

Established engineering, design,                None.  (I have NEVER in my
    manufacturing and QA/QC                       entire career seen a soft- 
	methodologies.                                ware QA/QC program that
												  had a hope of working.)

And considering that the price of cars has not kept up with the cost of
goods sold while software prices have been on a steady increase over
the last couple of years, arguing the cost of supporting a good warranty 
sounds even sillier.

>I suspect that a lot of the time, software vendors are just strongly
>interested in being able to charge enough for their products to recover
>the product costs within the lifetime of the product.


>    Software vendors are providing what is demanded of them.  It's not that
>    they set out to produce buggy software, nor that they're employing in-
>    competent engineers to build products.  (They keep ME away from ISC's
>    product, you'll note.:-)  But the overwhelming demand is NOT for
>    reliable, efficient software!  The demand is for features!  It's "give
>    us this and that and the other...and we'll buy it from the first vendor
>    who offers it!  We don't care if it's slow, or badly designed; we don't
>    care if it will give you maintenance headaches unto the seventh gener-
>    ation.  We want it all, and we want it NOW!"  The demand that the soft-
>    ware actually work is only made later, and only made of the vendor who
>    first satisfied the feature-list requirements and got the sale.  Fea-
>    tures requested may never be needed or used.
>If you accept that premise, you can see what sort of software it engenders.
>Should you accept the premise?  Let's kick that around for a bit.

I dont' accept that premise at all.  The most I've heard from users as
a software development manager has been that they'd like the provided 
features to actually work.  Not an unreasonable request.  Even most
new feature requests are really requests to get omissions in the original
product.

To the extent that we (customers) actually pay real money for broken products,
we do have to accept part of the blame.  No one would ever buy a car whose
engine would not run unless you held the little dohicky at a certain angle.
Yet, citing ISC's inode bug, the security bug, SCO's C2 debacle, almost
any binary version of sendmail and the list goes on and on, we sucker 
customers just keep buying and even recommending the products.  To an
extent, this is because we have no other choice.  It would bode vendors well
to remember that often times when consumers seem to have no choices, they
look to governments for fixes.  The array of automobile-related laws is
a sad example.  Does this industry really want that?

The premise that a vendor must charge large bucks to cover expenses before
the product obsoletes is equally empty.  The lifetime of most major
successful software is measured in fractions of a decade.  Unix - almost 20 
years.  Lotus - going on 10.  WordStar - going on 15.  WordPerfect - going
on 8.

I look around at the tools I use and with one or two exceptions, they
are the old familiar one.  Unix.  Wordstar. A Lotus clone.  Microsoft Word.
The one new tool I've bought recently, CorellDraw (WOW!) meets most of
my criteria for good software.  A fantastic product with no discovered bugs
yet, unlimited support, decent documentation including a instructive video
tape, and a fair price.  (not good but fair).

I've purchased upgrades to all these products in order to gain new 
features.  That's what the upgrade fee pays for.  Where the problem comes
in is when a vendor tries to justify many hundreds or thousands of 
dollars for a mature product by claiming development costs.  I wonder if 
any of the original Lotus people are even still there, much less being
paid?

I think several things are at work here to conspire to make bad, expensive
software happen.  First and foremost, there has crept in this concept that
software can somehow be engineered.  This implies that software can be 
produced by droids using engineering formulas/CASE/etc tools.  Part of
this stems, I think, from computer academia wanting to call sofware development
"engineering" so that they can have a "real" engineering program just like
the MechE and EEs.  They can't be satisfied with software development as a
creative enterprise more akin to art than engineering.  True, algorithms
can and are engineered; product development most assuredly cannot.

Again, with an automotive/manufacturing background, I find this 
extraordinarily strange.  The designers at an auto company consider themselves
artists who work in studios.  Same for the product developers at, say, a 
food company.  

Good software is invariably written by creative people passionately 
involved in the product.  Products written by droids who work 9-to-5 
jobs and who could care less what they are doing are the software 
equivalents to Elvis on black velvet paintings.

The second problem is that many software companies are run by people who
are in the business for all the wrong reasons.  Between those looking
for quick bucks with memories of when a CP/M floppy and a xeroxed manual
would bring $500 and those who are just flat incompetent, many companies
are a sad mess.  I've worked for some of the worst.  In many companies,
even the process of hiring programmers is bankrupt.  At my last company,
it caused the HR droids to spin in place when I actually proposed to TEST
prospective development programmers.  They went catatonic when I specified
that candidates would have to bring and present example software during
the interview.  I suppose that all I was supposed to do was to evaluate
their power ties.

Companies with genuinly good software are almost invariably lean and mean
and the developers are passionate about what they do.  Take my new favorite,
Corel.  The entire company including their consulting and laser disk divisions
is about 80 people.  Hell, at the last software company I worked at, there were
more than 80 people in the department that tried to work around all the bugs
for the customers.

>Software is actually worse, because of compatibility:  We're carrying every
>tailfin and chrome strip for the last twelve years.

Actually, in many cases, its the useless chrome and tailfins as excess
baggage that's the problem.  Let me pick on ISC again for a moment.  Look
at the useless tripe that helps make up the 60 lbs of books and disks.
LookingGlass?  LPI?  Topviews? (is that name correct?)  Most of this 
stuff is useless and even if it worked as advertised, I'd pity the 
company that developed solutions on top of these proprietary tools.
I have to wonder who came up with the idea to put this stuff in the
distribution.  I can't imagine customers beating the doors down demanding
a copy protected windows manager like LookingGlass or a compiler as pitiful
as the LPI stuff.  What we want are good, reliable hotrods.  Fast, efficient
and simple.

>Pick a pair of vendors in any software market--say Sun and HPollo, or AT&T
>and OSF, or Novell and Microsoft...doesn't matter.  Look at what they're
>advertising; look at how they compete.  Is it quality? performance? relia-
>bility?  NO.  The ads point to price, features, and delivery dates.  How
>much "stuff" can you get, for how few dollars, and how soon?

>They don't.  OK, let me stir it up again, with another contentious devil's-
>advocate (personal!) hypothesis:  An unsatisfied customer is better than
>none at all!  If people buy from you based on features, and only forsake
>you if you do a really bad job, then clearly the way to maximize your
>customer base (hence revenue) is to provide the minimal level of support
>required to keep from losing customers--let 'em grumble all they want, but
>stop 'em short of leaving--and put the rest of the effort into features.


This is not a devil's advocate hypothesis.  It is fact.  I have seen that
precise attitude in at least 2 companies I've worked for.  The software
business is now about where the automobile business was in 1968.  
The question is, who'se gonna be the Ralph Nadir of software and do 
we really want what such a cretin will produce?  Do we really want
a National Software Safety Administration?  I don't think so.

>Obviously I'm down on the feature frenzy...but let me belabor it just a
...
>This game is played on a
>daily basis in operating systems, and people are still falling for it!

Yeah, kinda like it would be in the auto business were it not for 
legally enforced SAE and DIN standards.  I hope it does not come to 
legally enforced standards in the software business.

>And if you agree, even in part, with my frustration at the feature wars,
>suggest how we get off...because I think it's a treadmill for vendors AND
>users.

I'm not sure I agree that feature wars are a major component in software
cost.  I'd more likely agree that it is a source of many bugs.  In any
event, the solution from the vendor's perspective involves strong 
participation in standards activities and selectively just saying NO
to creeping featurism.  Unnecessary embelishments on established standards
should receive the same level of distain as would a "new and improved" 
bolt thread standard.

John

-- 
John De Armond, WD4OQC        | "Purveyors of speed to the Trade"  (tm)
Rapid Deployment System, Inc. |  Home of the Nidgets (tm)
Marietta, Ga                  | 
{emory,uunet}!rsiatl!jgd      |"Politically InCorrect.. And damn proud of it