The FDA doesn't actually do testing. The device maker supplies evidence that development followed a process that includes testing, the types and amounts of testing being based on the risks posted by the device.
A company can potentially lie, and claim they did testing that they didn't - but Goddess help you if FDA figures that out: you're in deep, deep trouble. And they can definitely figure it out during a regular inspection, or if people get injured by your product, etc.
I'm working with some other folks to start a company to develop, manufacture, and market open-source medical devices. We all have extensive experience in developing commercial medical devices - defibrillators, radiation therapy for tumors, etc - and we're convinced that getting more eyeballs to review software and hardware will substantially increase safety and reduce costs.
Yes, we know how to work with the FDA and so forth.
It turns out that in managing batteries and booting there are a ton of oddball cases that cause things like this to happen. For example, there may be enough power available from the battery to start the boot process with the CPU in a low power, but once some peripherals start turning on the power draw bec omes more than the battery can support, a voltage rail drops to low, and a reboot happens. Shameless plug: there's a chapter in my book on product development that covers some of these issues and solutions, http://www.goodreads.com/book/...
I had Ubuntu running on our main home computer for more than a year. There were definitely some annoyances, but things were tolerable. Until I upgraded Ubuntu (I forget which version) and sound was broken hard - I spent hours fixing it, but got calls from home every few days asking me to "help fix the sound. it's broken again". Finally, I gave up and switched back to Windows.
and today Sanjay "I just make crap up about Michael Moore" Gupta, and the RIAA golden boy.
Obama is surrounding himself with some pretty interesting characters. Not good.
I registered a bug with OO 6.5 years ago, still unfixed, that causes spreadsheets to give utterly wrong results in even the simplest calculations. Sometimes OO treats a number as a string, and assigns it a value of "0" in calculations, e.g., 1+1 could equal 0 or 1.
Either OO should throw an error "can't treat a string as a number" or it should guess the number of the string is a valid number. But a major undetectable error like this is murderous, as has been testified to by the folks reporting the same bug after I did.
(Note the OO bug tracker seems to be having problems at this moment, so the link doesn't work.)
LiPo batteries (used in iPods and lots of other devices) are little roman candles when the go off: video here.
Some LiPos are pretty sketchy - we've had a few in prototypes "pillow" (fill with gas), but no explosions yet. Definitely a no-no in medical devices, particularly implantables!
Being sold at Walmart is *exactly* what the OLPC (or any similar product) needs.
"New" and "different" are lovely words to engineers, but they are terrifying to bureaucrats. I'm sure that the folks deciding whether to buy OLPCs are thinking "What if these folks disappear? I'm left holding thousands of units that may or may not be of any use" and they flee to the safety of Wintel - to paraphrase the old saw, nobody ever got fired for buying Wintel.
Instead of being silly and shunning sales to the non-poor of the world, the OLPC folks should be hawking it everywhere they can - Walmart, mail order, US schools - everywhere. Only when it becomes ubiquitous in the non-poor world will it take off in the poor world.
I've written a small online book on my experiences in managing technical projects.
One of the biggest mistakes I see is that managers don't recognize that there will absolutely be some failures within the project, i.e., approaches that turn out not to work. It's important, when one can, to move high-risk stuff to the beginning of a project, and even have a "pre-project" phase where the unknowns are explored and conquered, leading to a much better spec, and much better time and cost estimates.
Another important thing is to figure out what end-users actually want, not what they think and say they want. Sure, in a perfect world they could actually tell you what they want - but this ain't a perfect world. This involves a lot of watching users work and asking questions as they do their job.
Actually, folks over 65 in the US already do have single-payer insurance - Medicare. Turns out that once Americans make it to 65, their medical outcomes become similar to the rest of the industrialized world, presumably because their health care system starts to look like that in the rest of the industrialized world.
And in every case it costs far less (half as much, on average), and provides better medical outcomes (we're something like 38th in the world). Since it demonstrably costs less AND works better every time it's been tried, it's more like being the last to trade a horse for a car than it is like following the others off a bridge.
A few comments from this Massachusetts resident...
First off, this wasn't Romney's idea at all - the entire thing was proposed and implemented by the (extremely Democratic) state legislature. The MittFlopper had zero to do with it - absolutely nothing - he simply made sure to grab credit at the time (now he's distancing himself).
Personally, I think our country is jaw-droppingly stupid to not implement single-payer health care (aka Medicare for Everyone, aka What Almost All Other Industrialized Countries Do). That being said, the Massachusetts initiative has produced a number of very affordable plans, so I do think it's better than nothing.
But this "spectrum autism" label is quite misleading, and there's no evidence that it's helpful.
For example, kids who we used to call "nerds" are now labeled as having a "autistic spectrum disorder" (hence all of the interest from slashdot). One prominent psychologist that I spoke with said that many MIT professors have autism spectrum disorder - I'm not sure that "MIT professor" and "disorder" really belong together - these are obviously very-high-functioning people - I'm not sure that their inability to be used-car salesmen should be labeled a "disorder".
You're confusing "proof by repetition" with "science".
For example, pneumonia and lung cancer have various symptoms in common. So should we say that pneumonia is part of a "lung cancer spectrum"? Of course not - but this is precisely what psychologists are doing with a bunch of people who have certain characteristics in common.
One difficulty is that the psychology community keeps insisting that there is something called an "autism spectrum". Last time I did some research on this, I could not find a single piece of evidence to support a spectrum - in fact, the little evidence that existed indicated that there are several distinct conditions that have some symptoms in common.
Since the recent news tells us the White House can't even keep it's own email under control I can't imagine that they could defend against even an eleven-year-old script kiddie with a TRS-80.
Does it still have the 100% CPU utilization FEATURE that 1.5 has? I'd hate to lose the thrill of my system turning into a seized engine in quicksand... or the thrill of my less-computer-literate friends and relatives calling me for help on unfreezing their systems...
I have a fair bit of experience on the medical device side, not pharma - but they should be somewhat similar.
The FDA has standards for validationg COTS (Commercial Off-The-Shelf Software). ISO9000 is useful, but not necessary for FDA "stuff". For example, I don't think that any OS (Windows, Linux, etc..) is ISO9000 compliant, but your software runs on it, no? How could that be if ISO9000 was mandatory?
It seems to me that, for the FDA to be happy, you'd need to run some sort of reasonable validation of the COTS software, based on how it's being used; write a set of requirements, write a set of tests that prove that those requirements are met, and go to town. It would be good to recognize in the FMEA (Failure Modes and Effecta Analysis) that the COTS software probably has somewhat of a chance of failure - but then, the FDA (at least on the device side) has already announced that you must assume that software will sometimes fail, no matter how it's been tested. Finally, be exacting about audit trail - versions used, how things were built and installed, and so forth.
If your customer insists on third-party code being ISO 9000, despite the FDA not requiring it, then you're SOL. However, that requirement just doesn't make sense, unless they are not using an OS to run their software on (as described earlier).
I've actually run an effort to put software under ISO 9000 control - we passed the third-party audit, and we got a lot of software completed, and our product got approved by the FDA, and we followed our own rules - so I guess that it was a successful effort.
As you write, one key is to keep things lightweight, so that everyone can read and follow the procedures.
Another key is to write your own procedures - what usually happens is that someone snags a copy of another company's procedures and tries to use them. Other companies are different than yours. Do what works for your people and your product.
ISO9000 basically says three things: 1. do what you say you'll do 2. say what you actually do 3. don't let stuff fall between the cracks
1. and 3. are relevant to software development - 2. is aimed at the sales and marketing.
1. is a matter of having a process in place - having specs to work and test against, and so forth. Standard lifecycle stuff.
3. is best served by having a single place for everyone to enter any info that might affect the code - bugs, features, customer gripes, whatever. Once a week, have a cross-functional meeting - decision-makers from R&D, marketing, and whoever else is needed to MAKE DECISIONS. Then go through the list of issues, and make sure priorities are correct.
Having 2-3 pages of coding standards is also helpful. My personal favorite for the first rule is something like "It is the responsibility of those who write code to make the code's function quite clear to other programmers. All of the other standards are suggestions that may be bypassed for the sake of clarity of code - but think twice before you discard them."
Writing all of this up should take no more than 10-20 pages - try to make them somewhat "conversational", and use lots of flowcharts, so it's not too dry to read.
I assume that the algorithms and math are in Matlab. Matlab is much better than C++ for developing and unit testing math "stuff". However, shipping Matlab libraries with your application means a more complicated setup, licensing issues - and it will look pretty "bush league" (try it you'll see what I mean). Also, I'm guessing that your "domain experts" are more comfortable with Matlab than with C++ - which is why you're asking this question.
I would continue to develop algorithms in Matlab, and use the Matlab compiler to move the algorithms to C++ for integration with the C++ "presentation layer" code. Then compile and ship an all-C++ product.
My almost-six-year-old has had both for more than a year. The V.Smile is OK, but the Leapster is amazing. Using the Leapster, he's learned to read, add, subtract, and multiply. He's already facile with negative numbers - really incredible stuff for a kindergartener. And we haven't pushed him at all - he just loves the Leapster games that teach this stuff.
Since both products are relatively inexpensive, I'd recommend that you get both, and see how things work out.
The problem with your "if you don't like it, don't work there/shop there/vote for them" argument is that you ignore the reality that the wealthy have an absolute concentration of power. All corporations that we buy from and work for are, in the end, owned by the same group of predators that care only for themselves, and who will screw the working stiff in an instant if it gets 'em a new bauble from Tiffany. Both major political parties are wholly owned by the same folks.
Your observation that everyone thinks they're about to be a CEO (or otherwise a wealthy member of the predator class) is a correct observation, and another important issue - they don't want to hinder the gutting of the middle class for fear that it will hinder their own great money-grab.
Slashdot Mirror
The FDA doesn't actually do testing. The device maker supplies evidence that development followed a process that includes testing, the types and amounts of testing being based on the risks posted by the device.
A company can potentially lie, and claim they did testing that they didn't - but Goddess help you if FDA figures that out: you're in deep, deep trouble. And they can definitely figure it out during a regular inspection, or if people get injured by your product, etc.
Sorry to quibble... but medical devices are normally developed using a process that conforms to ISO 13485, not 9001. Pretty similar, though.
I'm working with some other folks to start a company to develop, manufacture, and market open-source medical devices. We all have extensive experience in developing commercial medical devices - defibrillators, radiation therapy for tumors, etc - and we're convinced that getting more eyeballs to review software and hardware will substantially increase safety and reduce costs.
Yes, we know how to work with the FDA and so forth.
Stay tuned...
It turns out that in managing batteries and booting there are a ton of oddball cases that cause things like this to happen. For example, there may be enough power available from the battery to start the boot process with the CPU in a low power, but once some peripherals start turning on the power draw bec omes more than the battery can support, a voltage rail drops to low, and a reboot happens. Shameless plug: there's a chapter in my book on product development that covers some of these issues and solutions, http://www.goodreads.com/book/...
I had Ubuntu running on our main home computer for more than a year. There were definitely some annoyances, but things were tolerable. Until I upgraded Ubuntu (I forget which version) and sound was broken hard - I spent hours fixing it, but got calls from home every few days asking me to "help fix the sound. it's broken again". Finally, I gave up and switched back to Windows.
It works and docs find it helpful. I'm amazed that it's ignored in TFA.
Docs won't use EMRs until they need to do so to get paid. That's the long and the short of it.
and today Sanjay "I just make crap up about Michael Moore" Gupta, and the RIAA golden boy. Obama is surrounding himself with some pretty interesting characters. Not good.
I registered a bug with OO 6.5 years ago, still unfixed, that causes spreadsheets to give utterly wrong results in even the simplest calculations. Sometimes OO treats a number as a string, and assigns it a value of "0" in calculations, e.g., 1+1 could equal 0 or 1.
Either OO should throw an error "can't treat a string as a number" or it should guess the number of the string is a valid number. But a major undetectable error like this is murderous, as has been testified to by the folks reporting the same bug after I did.
(Note the OO bug tracker seems to be having problems at this moment, so the link doesn't work.)
LiPo batteries (used in iPods and lots of other devices) are little roman candles when the go off: video here.
Some LiPos are pretty sketchy - we've had a few in prototypes "pillow" (fill with gas), but no explosions yet. Definitely a no-no in medical devices, particularly implantables!
Being sold at Walmart is *exactly* what the OLPC (or any similar product) needs.
"New" and "different" are lovely words to engineers, but they are terrifying to bureaucrats. I'm sure that the folks deciding whether to buy OLPCs are thinking "What if these folks disappear? I'm left holding thousands of units that may or may not be of any use" and they flee to the safety of Wintel - to paraphrase the old saw, nobody ever got fired for buying Wintel.
Instead of being silly and shunning sales to the non-poor of the world, the OLPC folks should be hawking it everywhere they can - Walmart, mail order, US schools - everywhere. Only when it becomes ubiquitous in the non-poor world will it take off in the poor world.
I've written a small online book on my experiences in managing technical projects.
One of the biggest mistakes I see is that managers don't recognize that there will absolutely be some failures within the project, i.e., approaches that turn out not to work. It's important, when one can, to move high-risk stuff to the beginning of a project, and even have a "pre-project" phase where the unknowns are explored and conquered, leading to a much better spec, and much better time and cost estimates.
Another important thing is to figure out what end-users actually want, not what they think and say they want. Sure, in a perfect world they could actually tell you what they want - but this ain't a perfect world. This involves a lot of watching users work and asking questions as they do their job.
Actually, folks over 65 in the US already do have single-payer insurance - Medicare. Turns out that once Americans make it to 65, their medical outcomes become similar to the rest of the industrialized world, presumably because their health care system starts to look like that in the rest of the industrialized world.
And in every case it costs far less (half as much, on average), and provides better medical outcomes (we're something like 38th in the world). Since it demonstrably costs less AND works better every time it's been tried, it's more like being the last to trade a horse for a car than it is like following the others off a bridge.
A few comments from this Massachusetts resident...
First off, this wasn't Romney's idea at all - the entire thing was proposed and implemented by the (extremely Democratic) state legislature. The MittFlopper had zero to do with it - absolutely nothing - he simply made sure to grab credit at the time (now he's distancing himself).
Personally, I think our country is jaw-droppingly stupid to not implement single-payer health care (aka Medicare for Everyone, aka What Almost All Other Industrialized Countries Do). That being said, the Massachusetts initiative has produced a number of very affordable plans, so I do think it's better than nothing.
But this "spectrum autism" label is quite misleading, and there's no evidence that it's helpful.
For example, kids who we used to call "nerds" are now labeled as having a "autistic spectrum disorder" (hence all of the interest from slashdot). One prominent psychologist that I spoke with said that many MIT professors have autism spectrum disorder - I'm not sure that "MIT professor" and "disorder" really belong together - these are obviously very-high-functioning people - I'm not sure that their inability to be used-car salesmen should be labeled a "disorder".
You're confusing "proof by repetition" with "science".
For example, pneumonia and lung cancer have various symptoms in common. So should we say that pneumonia is part of a "lung cancer spectrum"? Of course not - but this is precisely what psychologists are doing with a bunch of people who have certain characteristics in common.
One difficulty is that the psychology community keeps insisting that there is something called an "autism spectrum". Last time I did some research on this, I could not find a single piece of evidence to support a spectrum - in fact, the little evidence that existed indicated that there are several distinct conditions that have some symptoms in common.
Since the recent news tells us the White House can't even keep it's own email under control I can't imagine that they could defend against even an eleven-year-old script kiddie with a TRS-80.
Does it still have the 100% CPU utilization FEATURE that 1.5 has? I'd hate to lose the thrill of my system turning into a seized engine in quicksand... or the thrill of my less-computer-literate friends and relatives calling me for help on unfreezing their systems...
I have a fair bit of experience on the medical device side, not pharma - but they should be somewhat similar.
The FDA has standards for validationg COTS (Commercial Off-The-Shelf Software). ISO9000 is useful, but not necessary for FDA "stuff". For example, I don't think that any OS (Windows, Linux, etc..) is ISO9000 compliant, but your software runs on it, no? How could that be if ISO9000 was mandatory?
It seems to me that, for the FDA to be happy, you'd need to run some sort of reasonable validation of the COTS software, based on how it's being used; write a set of requirements, write a set of tests that prove that those requirements are met, and go to town. It would be good to recognize in the FMEA (Failure Modes and Effecta Analysis) that the COTS software probably has somewhat of a chance of failure - but then, the FDA (at least on the device side) has already announced that you must assume that software will sometimes fail, no matter how it's been tested. Finally, be exacting about audit trail - versions used, how things were built and installed, and so forth.
If your customer insists on third-party code being ISO 9000, despite the FDA not requiring it, then you're SOL. However, that requirement just doesn't make sense, unless they are not using an OS to run their software on (as described earlier).
Good luck!
I've actually run an effort to put software under ISO 9000 control - we passed the third-party audit, and we got a lot of software completed, and our product got approved by the FDA, and we followed our own rules - so I guess that it was a successful effort.
As you write, one key is to keep things lightweight, so that everyone can read and follow the procedures.
Another key is to write your own procedures - what usually happens is that someone snags a copy of another company's procedures and tries to use them. Other companies are different than yours. Do what works for your people and your product.
ISO9000 basically says three things:
1. do what you say you'll do
2. say what you actually do
3. don't let stuff fall between the cracks
1. and 3. are relevant to software development - 2. is aimed at the sales and marketing.
1. is a matter of having a process in place - having specs to work and test against, and so forth. Standard lifecycle stuff.
3. is best served by having a single place for everyone to enter any info that might affect the code - bugs, features, customer gripes, whatever. Once a week, have a cross-functional meeting - decision-makers from R&D, marketing, and whoever else is needed to MAKE DECISIONS. Then go through the list of issues, and make sure priorities are correct.
Having 2-3 pages of coding standards is also helpful. My personal favorite for the first rule is something like "It is the responsibility of those who write code to make the code's function quite clear to other programmers. All of the other standards are suggestions that may be bypassed for the sake of clarity of code - but think twice before you discard them."
Writing all of this up should take no more than 10-20 pages - try to make them somewhat "conversational", and use lots of flowcharts, so it's not too dry to read.
Good luck!
I assume that the algorithms and math are in Matlab. Matlab is much better than C++ for developing and unit testing math "stuff". However, shipping Matlab libraries with your application means a more complicated setup, licensing issues - and it will look pretty "bush league" (try it you'll see what I mean). Also, I'm guessing that your "domain experts" are more comfortable with Matlab than with C++ - which is why you're asking this question.
I would continue to develop algorithms in Matlab, and use the Matlab compiler to move the algorithms to C++ for integration with the C++ "presentation layer" code. Then compile and ship an all-C++ product.
My almost-six-year-old has had both for more than a year. The V.Smile is OK, but the Leapster is amazing. Using the Leapster, he's learned to read, add, subtract, and multiply. He's already facile with negative numbers - really incredible stuff for a kindergartener. And we haven't pushed him at all - he just loves the Leapster games that teach this stuff.
Since both products are relatively inexpensive, I'd recommend that you get both, and see how things work out.
The primacy of seniority in promotion unionized jobs is a fabrication of the Preator Class.
Looks like you've been listening too much right-wing media - try staying away from Limbaugh, Hannity, the NY Times, and so forth.
The problem with your "if you don't like it, don't work there/shop there/vote for them" argument is that you ignore the reality that the wealthy have an absolute concentration of power. All corporations that we buy from and work for are, in the end, owned by the same group of predators that care only for themselves, and who will screw the working stiff in an instant if it gets 'em a new bauble from Tiffany. Both major political parties are wholly owned by the same folks.
Your observation that everyone thinks they're about to be a CEO (or otherwise a wealthy member of the predator class) is a correct observation, and another important issue - they don't want to hinder the gutting of the middle class for fear that it will hinder their own great money-grab.
Feh.
Wake up, people!