Slashdot Mirror
Should Developers Be Liable For Their Code?
Glyn Moody writes "They might be, if a new European Commission consumer protection proposal, which suggests 'licensing should guarantee consumers the same basic rights as when they purchase a good: the right to get a product that works with fair commercial conditions,' becomes law. The idea of making Microsoft pay for the billions of dollars of damage caused by flaws in its products is certainly attractive, but where would this idea leave free software coders?"
if you get it for no price, you don't enjoy such priviledges.
If someone sells GPL based software, they are free to do so and pick up the tab on flaws in the product. Same goes for proprietary software.
This should have been done at least 10 years ago.
It takes a man to suffer ignorance and smile
Be yourself no matter what they say
If software controlling an aircraft crashes and causes the aircraft to crash too and that kills people, I'm pretty sure the software makers might end up liable too.
Actually it would probably be whoever decided that that software was OK to use in an aircraft. If I were to somehow get an aircraft and install Gentoo on some critical system, I'm pretty sure I'd be the one to get in trouble rather than the Gentoo or Linux (kernel) or Glibc people.
Except you just can't run anything for aircraft control. Read the fine print on software like Java, Windows and other items. You'll see it explicitly states you are not to use it for nuclear power plants, aircraft control and other life-critical applications. There are special rules for the super-critical stuff.
On the other hand, if this happens in a restaurant that calls itself high quality and advertises the famous chicken soup from a master chef and you get the same treatment, then there are numerous consumer protection agencies in Europe at least to fine the given restaurant.
That concept is so pathetic I don't know where to begin. Consumer protection agencies to fine a restaurant for poor quality and bad treatment? Are Europeans that big of pussies? What is wrong with "tell your friends they suck, don't eat there" and watch their business evaporate? You can't be serious that the government steps in for things like this!?
Learning HOW to think is more important than learning WHAT to think.
The idea that code should be perfect is a stupid idea: consumers don't want that. They want "good enough," not perfect. Perfect costs a great deal of money
Your comment is "insightful", but it is beside the point. This is exactly the same issue with all engineering. An object manufactured to better standards than needed for the purpose is an overly expensive object. The question rather is the web of responsibility. If Microsoft or Google or even somebody's shareware makes a claim of suitability, certainly the consumer should have redress when it proves unsuitable.
There are many other dimensions of this issue. For instance, the software industry is well known for adding pointless complexity - features that nobody ever asked for. If GE added a can opener to a toaster, they would be liable for any unexpected risks this reveals, but Microsoft can make Word so complex that businesses using it accrue large expenses related to training, etc., and risks related to misformatted and delayed documents and so forth - and yet Microsoft currently faces no significant market pressure from liabilities associated with having broken their own product.
Exactly. The problem with trying to enforce this kind of measure for software is that there is a cost/performance curve, and most people don't want to pay to be right up at the end of it.
Heck, no-one in the world knows how to get right up to the end of it. Even the guys at NASA, whose development process is awesomely effective at producing reliable software compared to the the commercial/home user industry, still get bugs. Given the nature of their work, their bugs can cost as much in a single mission as a bug in widely used home user software costs spread across the whole user base, potentially including a cost in human lives, so it's not like they're hiring stupid people or not trying to get everything perfect.
If you disagree, post your argument. (-1, Overrated) isn't your personal censorship tool for views you don't like.
Well, as somebody with an engineering degree, I know that we were taught that we were responsible for designs produced using software products. So, for example, if one used structural design software to design a building, and that software gave erroneous results, you are to blame, and not the software.
I make software that goes on an aircraft for a living.
All such software is required to be certified by the FAA, which has elaborate requirements for development, documentation, and testing (the applicable document is DO-178B).
I'm told that the reason for certification is not safety, but culpability. If your software satisfies the requirements and passes review by the FAA, then your company will not be held liable if it causes problems.
In essence, certification represents "best effort" engineering practices and tries very hard to eliminate bugs in the final product.
By the time a software package gets on a plane, many people have combed over it looking for problems, and the testers have spent a massive amount of time running it. There is a safety/failure hazard analysis which asks all the "what if" questions, and the flight crew has written procedures in case it fails.
If a bug is found after deployment (this happens occasionally) and it is discovered that there was a flaw in the certification process, all hell would break loose. It would open up the FAA and the company to all sorts of lawsuits from injured parties. The people who signed off on the certification would essentially be screwed.
The FAA is generally a bunch of bureaucrats. The one thing they do well is look out for their own interests.
Oh, and I worked for the company that got Microsoft Windows certified to run in the cockpit as a map display. It's Posix compliant, dontcha' know!