Slashdot Log In
Laws to Punish Insecure Software Vendors?
Posted by
CmdrTaco
on Wed Jan 16, 2002 11:34 AM
from the and-the-land-of-the-free dept.
from the and-the-land-of-the-free dept.
Gambit Thirty-Two writes "An influential body of researchers is calling on the US Government to draft laws that would punish software firms that do not do enough to make their products secure."
Yeah that'll work.
This discussion has been archived.
No new comments can be posted.
Laws to Punish Insecure Software Vendors?
|
Log In/Create an Account
| Top
| 581 comments
(Spill at 50!) | Index Only
| Search Discussion
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
(1)
|
2
(1)
|
2
open source (Score:5, Insightful)
Re:open source (Score:4, Insightful)
A critical point, I think. Keep in mind that these security holes are not exactly akin to a lock with a pink sticker that says "This lock doesn't actually work". A lot of research and experimentation is necessary in order to exploit those security holes. Research and experimentation carried out by criminals. As much as I would love to see software companies held accountable for the generally terrible state of software quality industry-wide, I'm not sure it's fair to hold Microsoft responible for making possible the actions of a malicious hacker. Is it Honda's fault a slimjim opens the door of my Civic?
A Certain Level (Score:5, Insightful)
> possible the actions of a malicious hacker. Is it Honda's fault a
> slimjim opens the door of my Civic?
Well, to get a realistic comparison, you'd need to compare on even ground. Pretend for a moment that your car door locks went to "locked" when you pushed the lock button, and "unlocked" when you pushed the unlock. However, they didn't actually engage the tumblers in the door, so when it's locked, the handle still opens the door. Now, there's a switch inside the door that you can get to by pulling the door side off, and when you throw it the tumblers connect and when the door says "locked" it now really means it.
Now, would you blame Honda if they didn't set the switch to "on" at the factory, and didn't tell anyone about the switch, and only acknowledged that it exists when someone in the field finds it and threatens to tell the general public?
I'd bet you would. That's a fairer comparison, and so yes, I think the companies that produce easily exploitable software should be forced to reckoning for it.
Virg
Re:open source (Score:5, Insightful)
It does not have to be that way. Why not put in exemption for software that comes with source code? The presumption could be that releasing source code allows the user to take responsibility for the correct operation of the software. Also consider that the OSS writer has little or no control over changes the user might make (and that's one of the main points, isn't it?)
Just like a LLP (Score:5, Interesting)
Re:open source (Score:5, Insightful)
That's a bit like saying a car company shouldn't be held responsible for putting faulty brakes on a car, since after all, the car owner could have replaced the brakes with something that worked.
Re:open source (Score:5, Insightful)
It does not have to be that way. Why not put in exemption for software that comes with source code? The presumption could be that releasing source code allows the user to take responsibility for the correct operation of the software. Also consider that the OSS writer has little or no control over changes the user might make (and that's one of the main points, isn't it?)
What needs to be made illegal are EULAs that absolve the software creator of guilt for flaws. Ford is liable for putting the wrong tires on SUVs and causing people to die. Ask Explorer owners (if you can talk to people that would buy one nowadays) how they would have reacted to such a license, and imagine how the courts would have reacted.
You've also made an excellent point about the futility of the GPL, but I digress.
How to track liability (Score:4, Insightful)
For instance, am I liable if I use the standard C function gets() in a program? I, as the program vendor, can argue that that's what was taught in my undergrad CS course, or I could point the finger at the language designer or C library vendor.
What about a program I write that communicates w/ other software via a standard protocol, and works perfectly if the other software adheres strictly to that protocol but fails in combination with another program which implemented that protocol incorrectly; am I to blame, or is the other vendor? What if the spec is vague?
As I've said in other posts, the potential for good legislation along these lines is there, but only with *heavy* involvement of people who understand issues such as these, along side of the industry lobbyists, consumer advocates and politicians.
Join the Libertarian Party (Score:3, Informative)
Everyone would be in violation (Score:5, Interesting)
Re:Everyone would be in violation (Score:5, Insightful)
Fine them? (Score:3, Funny)
Oh my, the irony (Score:4, Insightful)
It's always interesting when those who call for freedom and security for themselves can only figure out how to do it by reducing the freedom of others. Now they want to legislate software standards? Come on, you have to be against that.
Freedom of Speech (Score:4, Insightful)
An additional question would be should all software now come with a warrently that specifically disclaims the implied warrenty and states that there is no warrenty? Would it be legal under the proposal?
Re:Freedom of Speech (Score:4, Interesting)
Do you have the right of freedom of speech to utter other potentially hazardous comments? Yelling "FIRE!" in the middle of a crowded theatre is dangerous, and illegal. If you're engineering a bridge, does "freedom of speech" give you the right to design it so that it will collapse when people try to use it?
There is a wide legal history for freedom of speech ending when it causes harm to others.
Re:Freedom of Speech (Score:5, Insightful)
You don't need to open that whole kettle of worms at all, in this case. The right to say something does not equate with the right to sell it - unless it is sold for the purpose of communication (which commercial software is not.)
People who write software and then sit on it, or only give it to a few friends, cannot and should not be able to be held accountable for their software not working - unless (like yelling "FIRE!" in the middle of a crowded theatre) there is clear evidence of malicious intent (computer viruses.)
Someone who distributes software for free ought to be required to disclaim any warranties, which they allready do, and that is fine.
On the other hand, when you sell a piece of software there is an implied warranty of merchantability that you cannot disclaim. Extending that warranty to include security is not a free speech issue. Your right to write any code you want is still protected, you just cannot necesarilly sell it.
By extension, however, code written for the purpose of communication - including "here is how you write DeCSS" or the example code in a CS textbook - would still be protected, and you'd still have a right to sell it, whether or not it worked or was secure.
Be careful of what you wish for (Score:4, Insightful)
Seems to me this will have the least impact on those who need to pay attention to security the most(large software companies) while having the potential to make it harder for the "little guy" to write and publish software.
What about the click-thru EULA? (Score:3, Informative)
It always has a limit that anything bad that happens while using their product is not their fault.
Now IANAL but I thought that by clicking I Agree, that you were actually agreeing to that.
Boon to Corporate America (Score:5, Insightful)
Another good move for corporate America.
Microsoft is able to defend itself against the government. Are you?
Other Microsoft Failings... (Score:5, Funny)
For example Microsoft Bob.
I've been waiting for a service pack for it for years. I'm just not as comfortable hooking Bob up to the internet as I once was. Bob has gotten more viral infections than an old French Whore in a port town.
-Rothfuss
I agree (Sort of...) (Score:3, Informative)
Ford and GM shouldn't be allowed to produce cars that kill people, simply because they couldn't be bothered to make them safer - like exploding gas tanks - ok, so that's not such a great example... (grin)
But really, but the responsibility where it lies. If I put a system out on the net, and don't take some steps to make it secure, I should be liable for damages it causes when it's compromised. Same for SW companies. If you produce a product that doesn't meet the "reasonable" man test for care in producing the product, the maker should be liable for negligence.
I might go even further though, and add some criminal penalties too.
Software can be more reliable and bug-free and secure. (Go read the "Software Conspiaracy") Sure it will cost more, but what do you think all the virus outbreaks costs business and individuals. It's just a hidden tax. MS (and others) are just shifting the burden of producing software that works to the users. It's cheaper for MS to produce the software, but lots more expensive for the user to use them.
Finally, the legal system _IS_ part of the free market. The threat and actual loss of damages to a plaintiff balance the system of the market. It's not just buyers and sellers - and a wild wolly mess...
It just bugs me when "free market" proponents want to proclaim that the courts are unneccessary in the free market - bull! They are important and the market will not function correctly without them!
good concept (Score:3, Insightful)
A better solution is to allow people to sue software companies that produce software that does not do what it is supposed to do. For example, if Microsoft says they have the most secure servers on the market, they damn well better be that.
As soon as a few lawsuits are filed, things will change for the better. There's too much being "protected" by microsoft software for them to continue business-as-usual for long if they get sued for every nimda/code red/etc out there doing damage.
However, if the company puts out patches (such as through windowsupdate) and the user fails to apply them in a timely manner, it's the user that screwed the pooch, not the producer.
Before we decide this is such a great idea . . . (Score:5, Insightful)
Re:Bad Idea (Score:5, Insightful)
The Ford Pinto.
We have laws that tell auto manufacturers how they can build cars. Not in detail, no, but they have to meet certain standards or they just aren't legal to make. Note that business concerns don't enter into it. Making the Ford Pinto the way they did originally was a good business decision. It really did cost Ford less to pay out the death claims than to improve the car. It even arguably benefitted the consumers, because lower costs to Ford meant a lower price on the car and consumers were still buying them even after the problem became public so people obviously wanted them. The courts still held Ford criminally liable for building a car that blew up and killed people when they could easily have built one that didn't.
So why should we treat software any differently?
White Hats (Score:4, Informative)
Lawyers would start to be accused of Bugtraq chasing.
The report (Score:3, Informative)
Well, just in case you haven't the draft report is available for online perusal here [nap.edu]
PS I said NAS, not NSA. Just to be clear.
DMCA would nullify this! (Score:3, Insightful)
Um, yeah, that makes sense.
Legislation vs. Certification (Score:4, Interesting)
By charging higher premiums to insure companies using software with a bad track record, there are already market forces in place: include that difference in premiums in the TCO-calculations microsoft is so fond of to prove that Windows is cheaper than any competition, and make management aware of it (and make them wonder why that insurance company wants higher premiums for insuring against damages from security holes in that software).
Legislation could hurt many a small software maker, and it would also be subject to heavy lobbying from Microsoft to see to it that their interests are hurt the least, a better idea would be an independant (that's the hard part) organisation providing certification of software. Once that is established there could be legislation demanding minimum standards for software used in certain critic areas.
That way each software maker could choose how much to invest in security and QA, and it would be more transparent for customers how secure a product really is, so they wouldn't have to rely on the software-makers advertising for that kind of information. In effect the insurance conditions and premiums for different kinds of software are already an indicator for its security, and the insurance companies probably have a high interest in accurately estimating the risks, so probably they should play some part in ensuring the proposed organisations independance.
False Advertisement / Work as Advertised (Score:3, Insightful)
Second, if any laws are written, my guess is they would merely extend already existing more generic laws regarding false advertisement. Under such circumstances, software vendors would not be *required by law* to produce secure software. But, if their advertising campaign, sales representatives, software packages blatantly lead potential consumers to believe that their product is of "enterprise-level", "mission-critical-caliber", "secure", "reliable" or any such wording which implies "secure software", then the law could provide for some serious compensations to the harmed consumer.
To avoid endless legal battles over wording, the government should define an entity whose role would be to design, draft and maintain a *very specific* scale of security levels which defines strong standards for security features within software packages. The scale could not only provide very precise security requirements for software, but also standards type of compensation to the consumer for failure to meet each of its levels' standards.
Such scale should be massively advertised thru all media so consumers would know to look for a software package's rating on such scale before purchasing it for any mission-critical purpose.
We could let software vendors rate their own software packages according to this scale. If the scale is *specific-enough* and clearly defines levels of security, then consumers should have very strong cases to bring to class-action law-suits to seek compensation in the case such software should fail to meet all of the requirements defined by their advertised grade on the scale.
Such model would keep the government's involvment minimal and place all of the liabilities on the software vendor, so consumers don't ever have to seek compensation from some government-sanctioned entity which would assign ratings to software packages. We must keep in mind that computer software is by nature a highly volatile, constantly evolving, and rarely flawless type of product, as every new piece of software written is by nature "cutting-edge".
Unsafe at any speed (Score:5, Interesting)
It took legislation to make cars safe. The auto companies hated it. They fought every inch of the way. But it made the auto industry grow up and make their products really work, no matter what.
Every major industry goes through this transition, where society insists that the technology work safely. Railroads did. Steam boilers did. Autos did. Civil engineering did. Electric power did. It's time for computing to do it.
It's time for the software industry to grow up and stop hiding behind one-sided licensing agreements. Software is too important in modern life to be as crappy as it is.
Buggy Code == Fraud (Score:3, Insightful)
There should be criminal and civil penalties for withholding information about security risks. Right now I do not have the legal right to know about security risks that are discovered in systems I use, the creators of those systems are not legally required to inform me when a new risk is discovered. This means that I can not make an informed decision about how to protect myself from the problem. I can't even use a list of currently unresolved risks to help me decide what systems to use and/or purchase.
To me, the withholding of security risk information is a form of fraud. It is the same as rolling back the odometer on a used car. It is the same as selling Pintos with exploding gas tanks and the same as selling flammable pajamas to children. Companies must be required to release security risk information about their systems in a timely manner. They must be legally liable for damages that result from security issues between the time they discover the problem and the time they warn users of the problem. These kinds of penalties will force companies to create secure systems in the first place. And, to warn people in a timely manner so that they can take action to protect themselves. Although it is tempting I don't think the developers should be required to fix the system. But, a list of all outstanding security problems must be included in advertising and on the packaging of any system. People have to be able to make an informed decision about what systems to use. We put warning labels on beer and cigarettes, we require people to wear seat belts, we require the disclosure of the ingredients of all our food, we have lemon laws to protect us from unscrupulous car salesmen, and we have product liability laws that cover every physical thing we purchase. But, we have no equivalent legal protection from the purveyors of software snake oil.
The only way a company should be able to get out from under these penalties is to declare the product "dead", notify all customers of record that no more security support will be given for that product. Declaring the software dead should also require that the source code and/or system designs as well as any patent and copyrights to the system be released to the customers so that customers can arrange for other sources of security support for the system. At that point the company would not be allowed to sell, distribute, or accept any sort of payment including royalties and support payments for the software.
Stonewolf
This is not as far out as it firt seems. (Score:3, Insightful)
Consider, say, the hotel I was at years ago... they had an indoor pool. Before you used the pool, you had to sign a waiver... they had a stack of them in the pool room.
The waiver basically said using the pool was at your own risk, etc, etc.
Now... Dad asked his lawyer later, for kicks.
Say you drowned becuase you couldn't swim.. and they had no lifeguard. This document would protect them... it was fairly clear there was no lifeguard.
But.. say the diving board was in disrepair, and broke off while you were about to dive, causing you to fall and break leg... guess what? That contract doesn't absolve them of responsibility. Why? Because... it was reasonable to expect that the diving board worked.. the owner still had a duty to keep the area safe for it's users, regardless of their waiver. (If they wanted a waiver to protect them against that, they would have to clearly state the risks.. state that the facilities are in bad repair and broken.
Now.. software, we have these horrible EULAs... but still. I can understand how it's okay for a company to, say, protect itself from being sued over some little bug.. of COURSE they have to. Like.. say Excel crashes while you are in the middle of some work.. and you have to re-do it, so you are late for a meeting, so you lose the deal, etc.
Just as in the real world, where even a disclaimer can't generally release you of all obligation, so should it be with software. I don't know what the wording would be, or what would be fair... but software vendors should have a certain level of accountability for what they do.
Now.. how does this affect OSS? I don't know. Do I think OSS authors should be responsible for what they do? Yes, to a degree.. but there is a problem.. I don't think someone should be sued just because they shared some code with the world and it didn't work.
Unlike most, I read the report (Score:4, Informative)
It certainly does not claim that Microsoft is responsible for most security issues. If it had I would have expected Butler Lampson to have resigned from the board. It is not usual for NAS reports to target particular companies. It is not likely that David Clark would attack Butler in that way given that they are both LCS computing profs.
The statement about Microsoft is actually introduced from other sources but in such a way that the casual reader assumes it was a recomendation from the report. The only occurrence of the string 'Microsoft' in the text is Butler's accreditation.
Likewise I find it hard to find any recomendations. The majority of the report is simply a post 9-11 rehash of three previous reports by the same board. The nearest the report comes to suggesting legislation is:
Consider legislative responses to the failure of existing incentives to cause the market to respond adequately to the security challenge. Possible options include steps that would increase the exposure of software and system vendors and system operators to liability for system breaches and mandated reporting of security breaches that could threaten critical societal functions
That is quite a way from endorsing legislation, which is hardly surprising given the makeup of the panel.