EVERYONE TALKS ABOUT software quality, or the lack thereof, but who is going to do something about it?
That was the question many readers were pondering in their response to my recent column on the subject (see "
Battling the bugs
"). As you will recall, I opined that the time has come for software customers to demand bug-free products. Frankly, I
expected to be swamped by people proclaiming I've lost my few remaining marbles, but only a few came to that conclusion. There
were many differences of opinion on how to get there, but the great majority of readers seemed to be in total agreement that
bug eradication is a goal worth pursuing.
"Most experienced people agree that software bugs may never be completely eliminated, but most should certainly come close,"
wrote one reader. "Getting the latest release out to continue the cash flow is obviously more important [to software publishers]
than making a quality product. Most of us feel trapped in the cycle of upgrading hardware and software in order to remain
competitive."
More optimistic readers offered a variety of solutions to the problem of getting rid of bugs. Error prevention, total quality
management, and software-based flow sensors were among the development techniques endorsed by some. Several readers were of
the belief that software quality problems will disappear as software is increasingly delivered as a Web service. And, of course,
many argued that open source will be the ultimate answer to ridding the world of bugs.
Quite a number of readers pointed the finger at Microsoft, its operating systems, and the practices it has helped make standard.
"I do not believe that software can be made 100 percent bug-free, but the publishers should be able to produce something within
the 99.99 percent range," wrote another reader. "What irritates me is that when you find a bug, often the answer is that one
needs to either be on support to get the fix or one needs to purchase the upgrade! My wife once found a bug in a version of
PowerPoint where, if you were in 'notes' mode and saved your presentation, all of the pages of notes were deleted except for
the page you were on! Microsoft called that a 'confirmed problem' and told her to buy/upgrade to the next version. This is
total arrogance!"
Although I got a lot of guff over an analogy that a reader made to the auto industry in my last column ("Yeah, there's a
great role model for software companies to follow," one wag remarked), readers again held up other industries such as medical
equipment and circuit design as ones that know how to prevent serious defects. "With great wringing of hands and chanting
of dire predictions of $100,000 bug-free word processing programs, software developers declare that making software truly
reliable simply cannot be done economically," wrote a reader in the aerospace field. "We who build airplanes have been doing
it for a great many decades. Why? Well, what alternative is there? How many aircraft would a company sell if it guaranteed
that fully 85 percent of them would not crash or burn up in midair? Could it be that aircraft manufacturers are content not
always to recoup their capital investment in the first 90 days of a product's life?"
Most agreed that software quality can and should be better, but many argued that it's hard to know where the responsibility
for a bug truly lies. That's an important point, and one to which we'll need to devote another column.
But there were a few who disagreed. One reader at a major software company said I was out of touch with the real world.
"Bugs happen," he wrote. "We fix 90 percent of them. There are not enough people to fix all of them on a two-year cycle. It's
not [from lack of testing]. ... We know the bugs are there. We start with 30,000 and drop them to 2,000. When the users only
want upgrades four to five years apart and will spend the same dollars over again for a new version, which is mostly bug fixes
and (very little) new functionality, then they will have 99 percent of [bugs] fixed."
He's right -- that's a world I don't even want to be in touch with. Other industry professionals took a different view.
"The arguments claiming that software is too complex to ever truly be bug free ... are a red herring," wrote a self-identified
old gray-haired programmer. "The problem is really one of trade-offs, usually code quality traded for time/money. We have
the technology to create tight concise elegant solutions. We choose instead to produce bloated, bug-ridden, overly complex
stuff that gets to market 'inside the window of opportunity.' ... As to why these trade-offs are chosen, I tend to lean to
the lack-of-integrity explanation. Your buddy [Cem] Kaner is dead on."
Lots of talk, but who can do something about it? Another industry professional had the answer to that question: "The single
biggest factor contributing to The Great Bug Problem is the software-buying public," he wrote. "I'm sorry, but the buying
public as a whole will choose features over stability every time, even if they have absolutely no use for the features in
question.
And there you have it. The solution to The Great Bug Problem is not for software buyers to insist on more stable software,
but rather to put their money where their mouths are. Nothing will change until it is more lucrative to sell rock-solid software
than it is to sell software with more features than the competition."
All too true. But never has the time been better for software customers to demand a change. With the slump in IT spending,
the software vendors desperately want you to get back on the upgrade treadmill. First make them prove they can deliver the
quality you deserve. Everybody talks about software quality, but the only one who can actually do something about it is you.
E-mail
Printer Friendly
Reprints
(InfoWorld) - In the litigious world we live in, deploying a unified communications platform in your enterprise could...
