I did a bunch of work for a dot-com startup in the early 2000s focusing on vehicle-tracking applications. I have a daughter with Down syndrome; Downs kids tend to wander too, so we looked at this issue quite hard. The good news: the technology is pretty straightforward. The bad news: that's about the only good news.

Batteries
A GPS chipset enables a controller embedded in the shoes (or on a device strapped to the person) to know where it is. The second half of the problem is to transmit your location to somebody else. The simplest and cheapest approach is sending the data via the cell phone system--eight years ago we were using the digital control channels of the AMPS (analog) cell system; today you'd use G3. But think of the problems you have keeping your cell phone charged--how often would you recharge the batteries in your patient's (or your child's) shoes?

GPS
GPS is a really cool technology--but it is frequently viewed as the high-tech cure for what ails ya. It is not perfect. In particular, GPS depends upon an extremely weak signal--the GPS chipsets use DSPs to dig the signal out of the ether. GPS chipsets lose "lock" all the time. If the patient is wandering around outside in plain sight, his GPS coordinates are going to be accurate. But when the chipset loses "lock" on the satellites, tracking devices will continue to report the last known good position. This can be disastrous: the patient wanders from a nursing facility out onto the public street--and gets on a bus. Inside that nice, big aluminum box he can ride all the way downtown--and his GPS-enabled sneakers will continue to report that he's out in the nursing home parking lot.

There's a serious challenge to solving problems with technology--you also have to make sure that the people who depend upon that technology know (and act on the knowledge) that it must be maintained, or it will fail. Consider, for just a moment, how many people die of smoke inhalation every year even though they have smoke detectors in their homes. But they didn't change the batteries....

There's a much smarter solution
As I mentioned above, I looked at this issue long and hard with a dot-com startup eight years ago. As we looked at it, we found a substantially better solution than GPS. Project Lifesaver is a not-for-profit organization started in Chesapeake, Virginia that has developed a simple, effective solution targeted at Alzheimers patients, Downs kids, and other "wanderers." The patient has a small bracelet (like a hospital bracelet) attached to his or her wrist: once per minute the bracelet broadcasts a serial value on a digital (i.e. low-power) frequency. If/when a patient goes missing, the people responsible for the patient call the police or the sheriff's office. The cops arrive with two directional antennas tuned to the frequency: they go off in different directions, do a little bit of trigonometry, and Grandpa is back in the facility in less than ten minutes.

The Project Lifesaver solution is not perfect. They have the same battery issue that the "GPS sneakers" approach has (the GPS sneakers approach has been tried again, and again, and again). They also will only work with local law enforcement agencies--in our county the @##$%#^^# sheriff cannot be bothered. They have had a lot of success with local service clubs funding the cost of the bracelets, and (more important) paying for and replacing the batteries.

The GPS sneakers thing sounds like cool technology. Using differential antennas and good ol' trig is much more effective. My daughter still wanders off occasionally (and we live adjacent to a state park)--I wish we could take advantage of the Project Lifesaver program here.

Thanks for your response.

I don't mean to do something contrary to the spirit and ethos of SlashDot--but please allow me to apologize. You're entirely correct--my response to your post was really more of a generalized response to a number of posts I'd read that evinced an attitude of

• We're grad students, so
• We're smarter than you industry dolts, so
• We can't be held accountable for safety

That's not the point that you made, and I was unfair in teeing off on your post.

Australian standards?
But your response reminds me--you have been involved in defining safety standards for your lab. The practical effect of U.S. liability law is that, in essence, we don't care about safety standards: we are entirely focused on making sure nobody gets hurt. We cannot use "but my product met the safety standard" as a defense in court: the literature is full of examples of people who did stupid things with well-designed products and collected big damage awards when (surprise!) somebody got hurt. (Tractor-trailer driver pulls out onto a two-lane road, oncoming car cannot stop in time, driver of the car is decapitated. The family sues...the trailer manufacturer. And wins.)

Can you use "we met the safety standards" as an affirmative defense in Australia?

Oh, horseshit.

For the record, I'm a software developer. And over the course of my career I have done a lot of software development related to the production, distribution, and use of very hazardous gases and chemicals, including weapons-grade nerve gas agents; I presently work for a large engineering company that makes high-voltage components used on building construction. I am current on lab safety training for handling very high-voltage current (simulating lightning strikes) for UL certification.

First off, let's dispense with the silly notion that nobody is doing experimentation in industry. We're not just trotting out books of alchemy and chanting "bubble, bubble, toil, and trouble" as we stir the cauldron. We're figuring out hard subjects like how to build stuff without a) burning buildings down, or b) kill people. We evaluate issues like that all the time.

The big difference, as TFA plainly states, is that in industry you simply cannot ignore safety. Two of the hazmat producers I have done work for have much the same attitude toward safety: there is no such thing as an accident (corollary: there IS such a thing as a negligent attitude toward safety); and safety failure is inexcusable. (The former client I most respect assesses the cost of any safety violation to the manager with profit-and-loss responsibility for the business unit. He starts the year with a $0 line item for safety violations--he has to explain to the board of directors if he exceeds his line item budget. It tends to focus the manager's attention.)

Ignoring a safety issue (such as the citation of lack of safety equipment by UCLA's in-house safety folks three months before the accident) is manslaughter, pure and simple. The responsible parties should go to prison. The "business unit" (in this case, academic department) should be disbanded. Do that once, somewhere, publicly--you'll be amazed at the impact on everybody else. Somebody gets killed in the Chemistry lab? Fire the professor responsible for the lab, fire the department chair who allowed the professor to permit this kind of gross negligence, and fire everybody else involved. Then turn 'em over to the police.

Think that's unrealistic? It's pretty much the atmosphere in which a lot of industry functions. There used to be a time when a business could figure that an occasional death--and the resulting Workmen's Compensation claim--represented an acceptable cost of doing business. Those days are long gone--get somebody killed, and you can face criminal prosecution. And personally, I think that's a good thing.

Firstly, any software running in a medical device (i.e. mammogram machine) has to be certified.

Please re-read my post: I'm referring to the setup of a standard-issue PC with high-end video displays for the purpose of reading the mammography after the image was taken. No--it doesn't have to be "certified"--radiologists can read "film" on a notebook if they so choose.

FUD? Not.
I'm guessing that most of your study of the law has come from watching TV. The whole problem of "joint and several liability" in American tort law is that plaintiffs do not have to prove any "chain of causality"--they don't even have to prove beyond a reasonable doubt that the software you developed did anything wrong. (The "reasonable doubt" standard applies to criminal trials. The standard for civil litigation is demonstrating a "preponderence of evidence".)

In order to win a civil suit the plaintiff has to be able to demonstrate that an actual injury occurred; and that the defendants--jointly and severally, by a preponderence of evidence--caused that injury in whole or in part.

Examples:
Way too many to decide which one to use. Recall that, in my original post, I was the system architect for the claims management system used by a major commercial insurer for handling liability suits. Our design work involved reviewing hundreds of current and former liability claims--mostly those claims that went all the way to trial. I have subsequently been involved professionally in legal and financial systems in the U.S., Canada, and Asia.

Folks,

Posting angry comments here on SlashDot can be recreational--but all the ranting and raving anyone does here won't make a bit of difference in the real world.

What WILL make a difference in the real world, of course, is taking advantage of all of the links so helpfully provided in TFA. All you have to do is send a polite email to some of the people involved, pointing out that the two Loomis employees acted really foolishly; that the REI "loss prevention officer" made REI look...well, like losers; and that the Seattle Police Department really, really needs to send a couple of officers off to Constitution Camp.

Here's the email I just sent to the U.S. headquarters of Loomis (employer of the guards who started this nonsense):

Folks,
It looks like two of your employees went way, way far out of their way to find something to step in this morning in Seattle:

http://it.slashdot.org/article.pl?sid=09/05/12/2239211&art_pos=1

Follow the link: it shows a photo of your two employees, shortly before they grossly violated the civil rights of a law-abiding citizen. And had their stupidity compounded by the Seattle police.

Talk to somebody in your I.T. department who is knowledgeable about the Internet. Ask him or her what happens when a story like this gets posted all over the web. About how tens, or hundreds of thousands of emails flood in to the responsible parties (like, for instance, REI--YOUR CUSTOMER). And how that can have a really, really damaging impact on YOUR CUSTOMER's business.

Then you might consider the impact on your relationship with a corporate customer after two of your employees have exposed them to a phenomenal amount of really, really bad publicity.

You might want to think about how you could mollify this guy.

Cheers!

Oh--and just to save you the trouble, I emailed your corporate headquarters in Sweden to bring them up to speed on the story too.

Civil rights are like muscles. If you don't exercise them, they waste away.

I'm a bit mystified about this post--software developers already are liable for damages and/or injury caused by flaws in their work. There's a whole category of liability insurance (and matching case law to boot) around the subject. In the insurance business it's called "software errors and omissions."

If you're not aware of this...
you're either not in the software business, or you probably should be talking to your insurance agent. If you work in the United States, do business in the United States, or can be found to have a "business nexus" within the United States, you can find yourself named as a respondent in a U.S. liability lawsuit. And there's a nasty little element of U.S. liability law known as "joint and several liability" that essentially means that whether you are at fault or not, if the jury finds that the plaintiff was injured--and that some degree of the fault lies with any of the respondents, all of the respondents are jointly responsible to pay damages.

An earlier post in this topic wrote, in essence, "between the developers, the tester, the customer, the business analyst--good luck figuring out who made the mistake." That's the point of joint and several liability--they sue all of you. The jury doesn't have to decide whether it was the tester, the developer, the analyst, or the end user. So long as at least some part of the injury was caused by negligence of some kind--you can be found liable.

The Achilles heel of Open Source
Suppose you join a project on SourceForge--like a nifty project to develop Open Source Linux video drivers for high-end plasma video displays. You produce some really spectacular work--and you draw the applause of a small community of really high-end gamers when you ship your first release.

But, unbeknownst to you, a networking consultant in New Jersey finds your project, and uses it to provide a low(er)-cost solution for a radiologist who uses the same video card/monitor system to read CAT scans. The radiologist is reading mammograms.

(Ominous chord plays here.)

Seventeen months later, a woman is diagnosed with an "aggressive" form of breast cancer. Had it been detected earlier, she might not have required a mastectomy, or serious chemotherapy. Now she has lost both breasts, and all of her hair has fallen out. She--and her attorneys--want to know why the radiologist didn't find the problem in the mammogram seventeen months earlier.

Right. The mammogram the radiologist viewed on a high-end plasma display. Using an Open Source video driver. The one you helped to develop.

Another ominous chord. This time in a minor key.

You are in deep yogurt. And whether the video driver had anything to do with it at all, you can expect to be served notice that you have been sued in federal court. And you will then be staggered to discover just how much it costs just to respond to the lawsuit. And the fact that you didn't get paid a dime--hey, it was Open Source, right?--doesn't make a lick of difference.

Fairness, Justice, and the Law are Three Different Things...
Fifteen years ago I was the system architect on a project to manage liability insurance claims for a very, very large insurance company. A liability claim is a lawsuit--we evaluated all kinds of circumstances (prior history in this jurisdiction, prior history with this judge, who the plaintiff was, who plaintiff's counsel was, who our counsel was, yadda yadda yadda). We went through a bunch of factors, carefully weighing each of them, till we got to the end of the process. The very last questions were--does plaintiff have an injured child to show in the courtroom? Does plaintiff have a disfigured woman to show in the courtroom? Does plaintiff have a dying victim--particularly with soon-to-be-orphaned small children--to show in the courtroom? If so, then all bets were off--it did not matter in the slightest if our insured was at fault or not. The jury, invariably, was going to collectively say, "you just won the lottery" and law or justice wouldn't have much to say about it.

And that's what's going to happen to you...
As I wrote above, justice or fairness don't enter into the discussion. The jury, faced with a disfigured woman (and plaintiff's counsel will be sure to have sent her to a spa and a fashion photographer before she started chemo in order to be sure to document how beautiful she was before the double mastectomy and losing all of her hair) will feel a deep sense of sympathy. Women on the jury will insist that We Must Do Something--and who, on the jury, will argue, "hey--nobody said life is fair..." The jury will console any worry-warts with the knowledge that "the defendants are all insured" and they'll write the plaintiff a big check.

And here's where you get screwed...
The radiologist who read the mammogram but didn't find the cancer? She was insured. The networking consultant who assembled the system? He wasn't insured--but he is incorporated. And the corporation has no assets. You and your four friends that developed the video driver? Three of your colleagues are apartment-dwellers with no assets worth seizing. But you and the team leader are both homeowners.

The jury awards the victim $4 million. In the post-trial conference, the radiologist's insurer agrees to pay $2 million. You and your friend are on the hook for $1 million each. Fortunately, you have $1 million of liability coverage on your homeowner's policy. Unfortunately, your homeowner's policy doesn't provide liability coverage for your commercial activities. Your insurance agent, and the claims adjuster for your insurance company, explain to you that even though you didn't get paid a dime, by providing an alternative to a commercial product, you were engaged in commercial activity. They're going to disclaim--but, out of the goodness of their hearts, they'll make a "courtesy" payment of $25,000. You're on the hook for $975,000 in damages.

Then the forty other panic-stricken women who were given negative mammogram evals by that radiologist serve notice that they are suing you as well....

Understand clearly: this isn't FUD
This is simple, first-week-of-class insurance and business law in the United States. This is reality. And it is precisely issues like this that cause small businesses (and large businesses for that matter) to go bananas about "tort reform" any time the subject comes up. And the cost of this--the risk, plus the cost of buying insurance to hedge against this risk--is precisely why some software costs a ton of money. Way more than it should. It isn't just paying for overpaid developers, fat-cat executives, Ferraris for the sales guys, and Aeron chairs for everybody--it's also paying for whacking great insurance premiums and $650/hour lawyers.

If you don't know this, if you aren't prepared to deal with the consequences of this--even the consequences of somebody using your software for a purpose that you never even contemplated as a possibility--you shouldn't be playing the game. If you're writing software, for work, for pay, as a contractor, as a hobby, or as an FOSS purist on a mission to save the world from Microsoft--you had better carefully consider how you will respond in the event of a liability suit. Because "contributing to the community" is not going to spare you from liability for damages. That's not fair, but (as the old joke goes) nobody goes to Fair School.

Kutztown, Pennsylvania.

The first (and perhaps still only) municipality in the world to wire every single building in the town with fiber optic. As a municipal service. You pay for it the same way you pay for trash collection, sewer, water, phone, and electricity.

Hometown Utilicom (Kutztown's municipal utility).

Outsourcing has nothing to do with H1B. I'm talking about hiring staff on H1Bs at half the salary that you'd give to a US citizen.

I'm sorry that I wasn't more clear. What AIG did was to lay off Americans, and replace them with H1Bs working here, in America. This wasn't off-shoring--this was train-your-replacement-who-is-getting-one-third-what-you-do.

As far as I know, there has been no major disclosure or legal action against a practice like this. All that has been are stories that people have put up in the web.

Yes--there has. AIG (the company that got the biggest bailout) tried to lay off almost 90% of their I.T. people back in the early 1990s--and outsource all of their jobs to an Indian company. The backlash was fast and furious--Maurice Greenberg, then the CEO, spent major time in front of Congress explaining.

I'm not particularly afraid of H-1Bs (although they do depress wages)--but just want to point out that there is a factual basis for the complaints.

Agreed!

TextPad is terrific. Easy to use, easy to configure to your UI preferences, killer macro support (I find it invaluable for database scripting)--and it never, ever, ever breaks.

If only our banking system were this reliable....

Textpad.com