Interesting People mailing list archives

Software makers should pay for bad products [TECH UPDATE]


From: Dave Farber <dave () farber net>
Date: Wed, 03 Sep 2003 10:31:26 -0400

[ There is a long history of attempts to "certify" software and "underwriters labs" . The first one I knew was one SHARE created, then there was the ACM Algoriths etc etc. They all suffered from catching the simple problems that now a days most manufacturers catch. When you are dealing with parts that have to fit together -- like printer drivers and OSes and in dependant applications, all you will do with this proposal is to slow down innovation and create a large not very effective Good House Keeping seal like system. Sorry when a printer company flubs it more than once I go to another manufactory . If you hate MS - use unix or osx. I use all of them and each has a lot of problems and a lot of good parts.

Lets not panic , lets think!!

Dave

(not Dan :-)


=======================================


 DAN FARBER

Software makers should pay for bad products

Tony Scott is fed up with products that fail to work as advertised and with IT suppliers who duck responsibility for their products' defects. Scott, chief technology officer at General Motors, complains that buyers bear "almost 100 percent of the economic loss that occurs because of bad software." His solution: Underwriters Lab-type testing for software and technology insurance. Check out his modest proposals. > Read full story http://ct.com.com/click?q=32-Gx5IIrQC8HgvvROe3fUwZJmR.BsR

See also: Should Microsoft pay your security patch costs? http://ct.com.com/click?q=e1-6RvNQxk3PzzDOAZ0xdAE3pfi5qcR




[] COMMENTARY--There are two things that really bug me about buying information technology, and right up front I'll apologize if this sounds like something Andy Rooney might have written. First, I hate it when things that used to work really well are intentionally "improved" (meaning that they probably won't work) by well-established companies that should know better.

As an example, I recently bought a new entry-level printer for our youngest son, who is going away to college this fall. Installing a printer on a Windows XP laptop should be no more than a five-minute job. Not only did it take over an hour, but I'm still not sure how I eventually got the printer to work.

[]

Turns out that some genius at the not aforementioned printer maker decided that the tried-and-true installation software the company had used reliably for some time was no longer good enough. I just wish he or she had engineered something that worked and had first bothered to test it. Friends and others in the business often ask me for technology recommendations, and I'm now going to steer them clear of this entire product line--not because the printer is bad, but because it is going to be a maintenance nightmare until this gets fixed.

In the corporate world we experience similar challenges. But instead of wasting an hour, we can waste millions of dollars.

For example, before General Motors does a general internal release of our desktop software "build" for employees, we have to test it with hundreds of applications and at least a dozen hardware configurations. Most of the time things actually work pretty well. But there are always exceptions--and these cost us a lot of money in terms of extra engineering work. This can also represent millions of dollars in lost sales for a technology provider that gets scratched off our supplier list because of repeatedly bad experiences.

I'm convinced there should be the equivalent of the Underwriters Laboratories for software. (Editor's note: Underwriters Laboratories is an independent, nonprofit product-safety testing and certification organization.) The existence and use of such a service would drive a serious effort in the industry to use independent testing services. It also would guarantee that products meet some minimum standards. These standards should include the whole life cycle of support, from cradle to grave. With such a service, both consumers and businesses could have confidence that products would work reliably.

The second thing that drives me nuts is the lack of accountability for defects in technology products. The technology industry is arguably responsible for developing the tools that have enabled most of the great productivity gains of the last 20 years. But I also would argue that the industry is the greatest source of lost productivity as well. Think of the millions of worker-hours lost when machines reboot, programs crash, or password issues are being resolved-–just to name a few of the bigger drains on productivity.

Technology companies often have weak or nonexistent warranties for their products. As a consumer or even a large business entity, we have little recourse (except to vote with our wallets next time around). I'd like to see an insurance service developed in which businesses and consumers could buy defect insurance as an optional part of the technology purchase price.

When a product is discovered to be defective (including software), the insurance would pay for somebody to fix the problem or replace the defective product. It wouldn't take long for the insurance companies to figure out which companies make good products and which ones don't. In turn, this would drive the price of insurance either higher or lower, based on real data.

Right now, almost 100 percent of the economic loss that occurs because of bad software products is borne by the purchaser, not the technology company. A change in this situation would drive some real metrics into the process and finally force the technology industry to come to grips with an important issue--quality.

-------------------------------------
You are subscribed as interesting-people () lists elistx com
To manage your subscription, go to
 http://v2.listbox.com/member/?listname=ip

Archives at: http://www.interesting-people.org/archives/interesting-people/


Current thread: