Yep - I agree too.
There is an interesting and somewhat chilling (pun intended) article on this by Bruno Latour:
Has Critique Run out of Steam?
Latour writes:
"What has become of critique, I wonder, when an editorial in the New York Times contains the following quote?
Most scientists believe that [global] warming is caused largely by manmade pollutants that require strict regulation. Mr. Luntz [a Republican strategist] seems to acknowledge as much when he says that "the scientific debate is closing against us." His advice, however, is to emphasize that the evidence is not complete.
"Should the public come to believe that the scientific issues are settled," he writes, "their views about global warming will change accordingly. Therefore, you need to continue to make the lack of scientific certainty a primary issue."
Fancy that? An artificially maintained scientific controversy to favor a "brownlash" as Paul and Anne Ehrlich would say.
I'm not sure I'd call this the neoconservative mindset - but I don't have a good label for it. Basically it's in some people's interest to make the science appear to be problematic - this isn't just a lazy media habit of having to report both sides.
The rest of the article is very interesting and deals with the issue of how to look at the social forces on scientists (which is often viewed as attacks on scientists) while promoting and persuing empirically based knowledge.
Dinosaurs were very successful critters - they were dominant for millions of years. Mammals would never have expanded from their niche if the 'reset' button hadn't been hit at Chicxulub.
I'd be happy to own a product named after a dinosaur.
Years ago I worked with a bunch of economists in the US Federal Government - they categorized 'bugs' in their memos into three types:
Typos: Simple misspellings of words. Infrequent, easy to detect, easy to fix.
Writos: Incoherent sentences. More frequent, hard to detect, harder to fix.
Thinkos: Conceptually bonkers. Very frequent, subtle and hard to detect; almost impossible to fix.
Most 'late' bugs that I've seen in software projects belong in the last category - a lack of design or the failure to make a working mock-up leads to 'thinkos' which are only obvious when the application is nearly completed. These are expensive to fix.
Of course you're correct; you didn't read what I thought I wrote. Let me try again.
You can document, test, validate, and qualify your systems to a very high standard yet still have them fail in the field. Customers (mine and possibly yours) own the computers that software is running on. If *they* violate the implicit or explicit (in our case - *explicit*) instructions for operating the software it is difficult to guarantee that it will function to the level that they expect. Our instructions and legal contacts with customes sharply define who is responsible for what.
Concretely - if some P2P music filesharing software is installed on a hospital network - or if their site has been infected by a virus or worm - even if none of the machines running my software are directly effected - the peformance of the network may make the system inoperable.
Most hospitals don't let people install personal software on hospitals machines. Yet - I've seen it. Many hospitals permit machines at home to be used to remotely access systems. These are difficult to police and control.
I agree completely with your about part 11 compliance. My point is that part 11 isn't enough in the sense that it doesn't prevent people from misusing their networks and their machines by installing unvalidated software. A special case for 'installation' of unvalidated software are the worms and virii that nest in Microsoft operating systems.
True - most medical capture devices are isolated and safe. But as a developer who does work on 21CFR11-compliant systems that are delivered on Windows OS this won't save you on desktop machines for access to data (on servers you have a better chance - but it's not guaranteed).
For example: if you can use a web browser to access medical records and radiology images you've made this information much more broadly available in an institution. Most hospitals have VPNs - so a surgeon or neuroradiologist can log in from home to see if an emergency requires them to drive into the hospital or not. All you need is for that physcian's machine to have software installed by their kid off of the net and it's compromised. And once they log into the hospital over the VPN it can spread.
So - 21CFR11 is a wonderful thing for safety and prevents many sorts of disasters - but it's not perfect for preventing things like system DLLs being changed. You can do things like test the components that are loaded (and sign those components) but if you have a DoS attack where the only commmon intersection is TCP/IP it's hard to protect your customer.
And if you require them to use a special isolated machine then you've increased the cost of healthcare and either they will (a) buy a competing system or (b) not buy any system and that person with a stoke might not get a good diagnosis as quickly.
Some hospitals refuse to use IE or IIS on their machines because of past experience with viruses and worms. I wish more did. I know that these products aren't always the vector (well - in my experience they are) but it's odd that so little institutional learning is happening.
This is just a linguistic trick - an "undisclosed Fortune 500 company" is an oxymoron. All 500 of them have already been disclosed numerous times.
If SCO was really serious maybe they could show the top management of Undisclosed, Inc. with bags over their heads? Nah - that might look too much like a perp walk.
Who checks for that mapping? Someone in marketing, of course....
Sure. But the engineers and the regulatory officer of the company need to check too. My remarks were in the context of medical devices (where I make my living as an engineer).
In regulated industries marketing claims have a lot more weight than in market you describe.
I don't see a problem with this - you could have a design that wasn't completed yet was still useful for use (and sale).
FDA-regulated devices require an indication for use. This is essentially a set of marketing claims. It permits the user to make the clinical decision whether to use the product or not; it is also used during design and testing to validate that the device is what the user and regulatory agency think that it should be.
This is much more sensible than relying on design documents. What doctor would want to read a design document with UML diagrams or other hieroglyphic notations? Marketing claims are much clearer here.
Of course - the designers and managers need to understand the domain well enough so that they can vouch for a close mapping between the marketing claims and the product design.
Well - I'm not fond of Perl (although I do see its power) but I did hear this talk at OSCON and it was one of the most playful and thoughtful talks I've heard in a while. Not thoughtful as in George Steiner's musings on postmodernism - but thoughtful in that he was teasing and suprising the audience so that they were completely engaged. It was sort of like watching a magic act where rabbits were being pulled out of hats at unexpected angles. I think what you're missing in the written text is the timing and tone of voice that he used - sorry you weren't there. It was fun.
This explains a *lot* about perl. . I thought the same thing in two ways: (1) Perl is a motley and this shows why; (2) Perl needed someone like Wall for the community to form. Constructing both a language and community is more like performance art than an exercise in BNF. In general the audience enjoys the performance when the performer is also engaged - and I suspect he was having a blast.
If you like your philosophy written more seriously - please take some
Tristan Tzara
as an antidote.
I really don't see what extra powers the FBI needs here.
I think that the 'extra' is political legitimacy. Most people think that existing laws are for catching criminals and they don't see themselves as criminal. Once the FBI gets the 'extra' they will prosecute a few cases with a lot of publicity. It's just a tactic for moving the privacy/criminality boundary one step at a time.
Well - what about 'recaf' - decaffeinated coffee where one adds caffein tablets to restore the natural beauty of coffee?
I guess this would be like the naturally carbonated water (Perrier?) where they remove the natural carbonation and then restore it so that they are all the same.
It's a slippery slope - first you modify the coffee - then someone owns the idea of modifying coffee - then 'all your coffee belongs to us'. Be careful what you wish for.
There are other economic forces at work here. People are moving into the larger towns from the countryside; there is more trade with the outside world; there is also more crime (mostly theft) because many of the religious artifacts and species which are endangered in other areas now have economic value.
As the older (non-monetized) economy is disappearing there are many changes in people's roles and in the social hierarchy. Older political hierarchies are also changing as the King is moving the country (with much skill) toward democracy.
It's not just TV.
I wish them the best of luck; they are going to need it to keep their bearings in a more globalized world.
It's not like they have a lot of choice though. At the end of WWII there were three Buddhist kingdoms: Tibet, Sikkim, and Bhutan. Tibet has been absorbed by China; Sikkim was absorbed by India. Bhutan is the last one. If they are going to stay independent they need friends; and to have friends means that they need to trade with the outside world. It's a very special place - I hope that they can keep most of their culture while remaining independent.
The reason this article might not reflect what actually happens is that hydrogen production might be done on a decentralized local scale.
This might make leakage worse. When Mao decided to decentralize steel production the quality was very low and the environmental effects were miserable. It really could go one way or the other - this is probably more of a social design issue than a technical one. In any case you'll need economic incentives to keep leaks low (both the economic loss of the fuel and possible fines on leakers).
So yeah, maybe it will die out. But the question really is should we care?
I wonder if there are other cognitive skills that are taught with cursive writing. If playing video games helps develop visual skills surely cursive writing teaches fine motor control.
I only have sketchy knowledge of this - but Intermezzo is designed as a distributed file system whereas Braam described Lustre as a cluster file system. The main difference between them is that Lustre was designed to be much more scalable in the following ways: the distributed lock manager is more sophisticated and the I/O load balancing is done very cleverly.
I believe that Peter Braam (one of the authors of InterMezzo) has been working on a new file system named Lustre that builds on top of it's ideas. It's out in beta There is a detailed document here. Braam gave a nice talk on it at LinuxWorld this last January. I haven't tried it yet.
Slashdot Mirror
#5 Pterosaurs
Latour writes: "What has become of critique, I wonder, when an editorial in the New York Times contains the following quote? Most scientists believe that [global] warming is caused largely by manmade pollutants that require strict regulation. Mr. Luntz [a Republican strategist] seems to acknowledge as much when he says that "the scientific debate is closing against us." His advice, however, is to emphasize that the evidence is not complete. "Should the public come to believe that the scientific issues are settled," he writes, "their views about global warming will change accordingly. Therefore, you need to continue to make the lack of scientific certainty a primary issue." Fancy that? An artificially maintained scientific controversy to favor a "brownlash" as Paul and Anne Ehrlich would say.
I'm not sure I'd call this the neoconservative mindset - but I don't have a good label for it. Basically it's in some people's interest to make the science appear to be problematic - this isn't just a lazy media habit of having to report both sides.
The rest of the article is very interesting and deals with the issue of how to look at the social forces on scientists (which is often viewed as attacks on scientists) while promoting and persuing empirically based knowledge.
Dinosaurs were very successful critters - they were dominant for millions of years. Mammals would never have expanded from their niche if the 'reset' button hadn't been hit at Chicxulub. I'd be happy to own a product named after a dinosaur.
Typos: Simple misspellings of words. Infrequent, easy to detect, easy to fix.
Writos: Incoherent sentences. More frequent, hard to detect, harder to fix.
Thinkos: Conceptually bonkers. Very frequent, subtle and hard to detect; almost impossible to fix.
Most 'late' bugs that I've seen in software projects belong in the last category - a lack of design or the failure to make a working mock-up leads to 'thinkos' which are only obvious when the application is nearly completed. These are expensive to fix.
The Register claims that contents may have settled during shipping.
You can document, test, validate, and qualify your systems to a very high standard yet still have them fail in the field. Customers (mine and possibly yours) own the computers that software is running on. If *they* violate the implicit or explicit (in our case - *explicit*) instructions for operating the software it is difficult to guarantee that it will function to the level that they expect. Our instructions and legal contacts with customes sharply define who is responsible for what.
Concretely - if some P2P music filesharing software is installed on a hospital network - or if their site has been infected by a virus or worm - even if none of the machines running my software are directly effected - the peformance of the network may make the system inoperable.
Most hospitals don't let people install personal software on hospitals machines. Yet - I've seen it. Many hospitals permit machines at home to be used to remotely access systems. These are difficult to police and control.
I agree completely with your about part 11 compliance. My point is that part 11 isn't enough in the sense that it doesn't prevent people from misusing their networks and their machines by installing unvalidated software. A special case for 'installation' of unvalidated software are the worms and virii that nest in Microsoft operating systems.
For example: if you can use a web browser to access medical records and radiology images you've made this information much more broadly available in an institution. Most hospitals have VPNs - so a surgeon or neuroradiologist can log in from home to see if an emergency requires them to drive into the hospital or not. All you need is for that physcian's machine to have software installed by their kid off of the net and it's compromised. And once they log into the hospital over the VPN it can spread.
So - 21CFR11 is a wonderful thing for safety and prevents many sorts of disasters - but it's not perfect for preventing things like system DLLs being changed. You can do things like test the components that are loaded (and sign those components) but if you have a DoS attack where the only commmon intersection is TCP/IP it's hard to protect your customer.
And if you require them to use a special isolated machine then you've increased the cost of healthcare and either they will (a) buy a competing system or (b) not buy any system and that person with a stoke might not get a good diagnosis as quickly.
Some hospitals refuse to use IE or IIS on their machines because of past experience with viruses and worms. I wish more did. I know that these products aren't always the vector (well - in my experience they are) but it's odd that so little institutional learning is happening.
If SCO was really serious maybe they could show the top management of Undisclosed, Inc. with bags over their heads? Nah - that might look too much like a perp walk.
Sure. But the engineers and the regulatory officer of the company need to check too. My remarks were in the context of medical devices (where I make my living as an engineer).
In regulated industries marketing claims have a lot more weight than in market you describe.
FDA-regulated devices require an indication for use. This is essentially a set of marketing claims. It permits the user to make the clinical decision whether to use the product or not; it is also used during design and testing to validate that the device is what the user and regulatory agency think that it should be.
This is much more sensible than relying on design documents. What doctor would want to read a design document with UML diagrams or other hieroglyphic notations? Marketing claims are much clearer here.
Of course - the designers and managers need to understand the domain well enough so that they can vouch for a close mapping between the marketing claims and the product design.
You're solving the problem of where marginal ethics equals marginal politics - it's where greed is optimized. It's just standard micro theory.
This explains a *lot* about perl. . I thought the same thing in two ways: (1) Perl is a motley and this shows why; (2) Perl needed someone like Wall for the community to form. Constructing both a language and community is more like performance art than an exercise in BNF. In general the audience enjoys the performance when the performer is also engaged - and I suspect he was having a blast.
If you like your philosophy written more seriously - please take some Tristan Tzara as an antidote.
I think that the 'extra' is political legitimacy. Most people think that existing laws are for catching criminals and they don't see themselves as criminal. Once the FBI gets the 'extra' they will prosecute a few cases with a lot of publicity. It's just a tactic for moving the privacy/criminality boundary one step at a time.
You're correct - I was confusing this with Bt. Thanks.
It's not a question of 'jumping' - just Natural Selection (tm) at work.
Well - when the insects also become roundup-resistent then I suspect we'll lose a lot of soybeans.
It's a slippery slope - first you modify the coffee - then someone owns the idea of modifying coffee - then 'all your coffee belongs to us'. Be careful what you wish for.
As the older (non-monetized) economy is disappearing there are many changes in people's roles and in the social hierarchy. Older political hierarchies are also changing as the King is moving the country (with much skill) toward democracy.
It's not just TV.
I wish them the best of luck; they are going to need it to keep their bearings in a more globalized world.
It's not like they have a lot of choice though. At the end of WWII there were three Buddhist kingdoms: Tibet, Sikkim, and Bhutan. Tibet has been absorbed by China; Sikkim was absorbed by India. Bhutan is the last one. If they are going to stay independent they need friends; and to have friends means that they need to trade with the outside world. It's a very special place - I hope that they can keep most of their culture while remaining independent.
The reason this article might not reflect what actually happens is that hydrogen production might be done on a decentralized local scale.
This might make leakage worse. When Mao decided to decentralize steel production the quality was very low and the environmental effects were miserable. It really could go one way or the other - this is probably more of a social design issue than a technical one. In any case you'll need economic incentives to keep leaks low (both the economic loss of the fuel and possible fines on leakers).
So yeah, maybe it will die out. But the question really is should we care?
I wonder if there are other cognitive skills that are taught with cursive writing. If playing video games helps develop visual skills surely cursive writing teaches fine motor control.
I only have sketchy knowledge of this - but Intermezzo is designed as a distributed file system whereas Braam described Lustre as a cluster file system. The main difference between them is that Lustre was designed to be much more scalable in the following ways: the distributed lock manager is more sophisticated and the I/O load balancing is done very cleverly.
I believe that Peter Braam (one of the authors of InterMezzo) has been working on a new file system named Lustre that builds on top of it's ideas. It's out in beta There is a detailed document here. Braam gave a nice talk on it at LinuxWorld this last January. I haven't tried it yet.
Do you have any references to the studies about Scandanavian medical data? I'd be interested. Thanks.