Software Approvals For Consumer Markets?

Odkin asks: "Some friends and I are struggling with a hardware project which is stalled due to costly consumer market approvals (which is alright I guess). But it struck me, why are there only market approvals for hardware and not software? The hardware approvals include functionality tests that ensure that the product works as intended in any way the user would handle it (even unsuitable use). Would such approvals for commercial software improve the quality of the products, including minimizing the risk of data loss and heightening the security? In other words, would it facilitate or inhibit the creation of good software?"

Comments

EULAs and No Programmer Liability

[Anonymous Coward]

Pretty much becuase Programmers aren't considered profesionals like "Doctors" and "Lawyers" by the govts, as they can teach themselvfes, and hence are exempt from Malpractice.

Also teh EULA especially in UCITA states shields the software company from damages. Go read just about any EULA when it talks about damages if you don't believe me.

OK - I'll bite - what hoops are in your way?

[rcpitt]

If you're talking about UL and/or CSA or whatever then you're probably designing things wrong - this is why many such systems use a "wall-wart" power brick since that is all that needs to be certified if the power it puts out is less than 48Volts.

If you are trying to get liability insurance, that's another thing - you can spend as much money as you have and it may not help.

So... give us a fer instance on what you're trying to do - your box looks pretty innocuous.

Approvals are for a different purpose.

[HotNeedleOfInquiry]

Approvals are for specific safety and RF interference issues, not for functionality or reliability.

The FCC/CE wants to make sure that your widget doesn't interfere with the other widgets. UL/CSA wants to make sure your widget doesn't burn the house down.

I know that CE has some EMI susceptablilty stuff that isn't exactly safety, but for the most part, the issue is making a safe, non-interfering widget. The widget could fail in 2 days, as long as it fails safely.

You are posing a question that is pretty much unrelated to hardware approvals.

Re:Approvals are for a different purpose.

[wfberg]

So what you're saying is that you've never used a piece of software (widget) that interferes with another piece of software (widget).

Check the bottom of your keyboard. Next to the sticky notes with passwords, you'll find a blurb like " This device complies with FCC Rules Part 15. Operation is subject to the following two conditions:
This device may not cause harmful interference.
This device must accept any interference received, including interference that may causeundesired operation.
This equipment has been tested and found to comply with the limits for a Class B digital device, pursuant to Part 15 of the FCC Rules."

Note that, yes, it may cause interference, as long as it's not harmful. Of course, in an Intensive Care Unit in a hospital, or in a precision laboratory, interference is more likely to be harmful than in the desert. So they've specified the environment in which this should be the case; the blurb on my keyboard specifically says "office equipment", which is what Class B digital devices in Part 15 of the FCC Rules are all about..

Of course, though equipment is tested extensively, this does not mean that there is zero chance of it ever generating interference. So, in the real world, we know this, and there are certain thresholds of interference under which (all, and thus, also the surrounding) office equipment must accept interference as well. Yes, an EMP weapon will fudge up your keyboard, but a cell phone shorting out next to it won't.

Random testing is a software testing technique which is appropriate to this kind of probablistic approach to understanding failures.

Re:Is software a bridge or a burger?

[Anonymous Coward]

Speaking of bridges...

http://www.cs.bell-labs.com/cm/cs/pearls/sec073.ht ml

I think every software developer should read that... perhaps annually. :-)

Re:Video game makers do it

[Jason1729]

Indirectly, that's why I do most of my gaming on a console machine. The quality of most PC based games is horrible. In Bridge Commander, it always freezes at exactly the same point near the end of the game (and that's on a few totally different machines since the game came out). I can't play Lords of the Realm at all anymore because for the past few years it detects my original CD as a pirate copy. Except for Blizzard's products (which always seem to be good quality), it seems like most of the PC games I've tried in the past few years just don't work. I've never had a problem with a console game.

Jason
ProfQuotes

Aircraft software

[old_unicorn]

Safety critical software such as aircraft software has to be highly checked and certified. The problem with software which does not happen with hardware is that a a change which is supposed to only fix problem (a) means a recompilation, and there is always the chance that function (b) has been screwed up, so generally we have to restest just about everything, whatever the change. It costs much more to test software to a good standard, than to develop it. We reckon that the minimum change for engine control software will take 6 weeks of 18 hours per day testing to validate. And we DO find unrelated faults, which occur occasionally.

I am in the process of doing this now.

[Dead_Smiley]

I am a Controls Engineer. One of my duties is to write software for automatated manufacturing equipment. I am going to have to do FDA validation on my software for functionality, mostly for faults and fault recovery.

my $.02

Government regulations, Lawsuits, and Free Markets

[R2.0]

Hardware testing is done because there are a number of Government regulations that require it (FCC, CPSC), product liability requires it, and the common law treats hardware just like any other property. After those minimums are met, the free market kicks in, driving toward better reliability/lower cost/pretty design - whatever the market wants at a given moment.

Software is a new animal, and neither the government nor the common law has caught up with it. (Hollerith cards @ 100 y/o vs. the Code of Hammurabi @ 1000's y/o.) As a result, software makers have had free reign in the market, because there is no mechanism that sets minimum standards. Seemingly absurd licensing practices are not challenged because there are few on-point cases in the common law and our legislators simply don't have the mental prowess to see software as a different class of stuff - part real property, part intellectual property, part printed word, part device. Nor do they have any incentive to work the problem out.

One of 2 things must happen to get software on the same footing as hardware:

1) Legislative action addressing the fundamental nature of software and how the law will treat it (I personally favor killing both software copyright AND patents and coming up with a 3rd classification)

2) Bold precedents in the common law to extend existing legal concepts to the current situation - unconscionable clauses applied to EULA's, detrimental reliance, or recongizing tort claims by users against software makers.

Either one will take someone, either judge or legislator, with some real balls. Other than that, don't hold your breath.

Re:Four words

[AKAImBatman]

Designed for Microsoft Windows XP

The only software I've ever seen with that label has been from Microsoft. There's plenty of hardware with that label, but when it comes to software, Microsoft doesn't share. Besides that, there are no real standards to live up to. The Nintendo Seal of Quality wasn't just "It uses the newest APIs". It was a actual scoring of a product done by human beings. If the product failed to meet Nintendo's standards (for whatever reasons, some perhaps made up on the spot to handle the differences in the game) it would have to be fixed or the product couldn't ship. I don't see Microsoft preventing anyone from shipping anything.

Already there ... Software Approvals

[Anonymous Coward]

If you build safety critical applications for a living, you are already aware of the hoops you have to go through to get the software approved for use. Definition of "safety critical" for the uninformed: You screw up, it gets through ... people die. Remembering that helps to keep things in perspective and helps to ensure that you develop software the *right* way even if it isn't the cheapest.

Weapon systems S/W -- DOD approval

Commercial Avionics S/W -- FAA approval

Non-military Nuclear Systems S/W -- DOE and other agency approvals

Manned Spaceflight S/W -- NASA and I mean *everybody* at NASA (or so it seems)

Software used as a medical device...

[nucleon]

is strictly regulated by the FDA. Not only is a software company required by law to obtain premarket approval 510(k) from the FDA before marketing certain types of medical software in the US, but it is also required by law to document and follow a very thorough software development and validation process.

Although this kind of software is usually not sold to the general public, it is used every day in hospitals and clinics to do everything from analyzing bacterial infections to robotic surgery to radiation oncology treatment planning.

I have worked for several software companies, developing software that is considered a class II medical device. Not only did we have to completely document everything from requirements to validation testing, but we had to keep the documents themselves under version control! Knowing that your software could mean life or death to someone, really puts the software engineering process into perspective.

Re:Four words

[mwood]

Microsoft shares its Logo Requirements just fine. I wish somebody would read them sometime. It's probably the finest thing to come out of Microsoft. Unfortunately even some of their *own* developers seem to have difficulty sticking to a few simple principles.

You don't see Microsoft preventing the shipment of shlock software, but you also don't see the shlock shipping with their holy Logo on the box. They can sue you for that, and don't think they wouldn't.

(Alas, universal acceptance of the Logo Requirements wouldn't help me personally since it simply means something was designed for an OS other than the one that best suits my way of working. But at least I'd lose less hair setting things up for the people who do want it.)