Any/. parents (or loco parents / aunts / uncles / cousins etc.) care to discuss good age ranges for introducing kids to Mindstorms ? It's very easy to hot-house kids into small-parts Lego when they simply don't have the motor skills to assemble the more fiddly parts.
Ignoring the moral argument for a moment (that unfit or disabled people are just as "deserving" as the fit), then imagine what would happen to the software industry if guys had their code junked because they were too fat ! How many really excellent coders do you know who are (by any reasonable measure) grossly overweight ? Some of the best people around for many jobs are also some of the most unfit, and it's a foolish employer that discards good workers on an irrelevancy.
It is rather unfortunate that the BBC correspondent has very little idea about the subject he is writing about
It's always the same from the Beeb. "Real" intellectuals have arts or humanities degrees, mathematicians are just geeks and beneath contempt. Did you notice the "related links" they placed on the page ? The top feature was last year's "mathematics of biscuit dunking" story. This just shows what little significance the increasingly dumbed-down BBC now places on science or technology stories.
ObRant: Why is it that at the hypothetical mixed-background middle class dinner party, the scientists are expected to be literate, but the literati still revel in their innumeracy ?
It is true that the subject is too esoteric to be accessible to non-mathemticians,
There's probably at least a dozen people in this room as me who work on elliptic curves on a daily basis. OK, so I work in an unusual environment, but these things do actually have real world applications (crypto, natch) and not just for the NSA.
Insurance is a complex business, and prone to many forms of fraud. To prevent this, most countries require that "insurers" comply with some pretty strict (sic) legislation before trading as an insurer. If you're a big company, you can maybe afford to carry your own insurance (and it probably makes sense to do so), but it's likely to be quite illegal to do so unless you also comply with the rules for "being an insurance company".
As an obvious example, consider UPS not delivering something (their fault) and then arguing that it really was delivered (your fault). If they also suffer the financial loss as the insurer, would you really trust them to admit that they'd lost the package and so should pay up ? I'm only a little guy - no way can I (usually) afford to sue huge UPS over it, so being reluctant to pay legitimate claims would make good economic sense for them (not that I claim they'd ever have done this).
Same as ever. Your working style is as young as your mortgage, partner and kids will permit.
"To live free and happily, one must sacrifice boredom. It is not always an easy sacrifice." Richard Bach, I think (Which I guess qualifies me as an old fart !)
The killer for Cray was when large custom boxes full of cheap commodity processors started to appear (Connection Machine et al.). Cray's previous expertise of making hugely clever processors just couldn't compete with Fordian economics and a huge fab churning out Intel games boxes by the bucketload.
Now Beowulf takes it all a step further and replaces a weird box full of standard CPUs with a room full of standard boxes.
Another (quite reasonable) opinion is that Cray was/b> Seymour Cray, and without him they just lost direction.
If there's no "source", how can you be expected to write self-modifying executables ? 8-)
Seriously, I think there's always "source", even for a Lego RCX or Seymour Cray toggling bit patterns into a front panel. Maybe you wrote the executable beforehand, but there has to be something that would pass as a "listing", so that others could replicate it by re-typing what you did.
Darwinian evolution is wrong in big chunks anyway, no-one except a few badly informed scientific bigots still claim it's all right.
It's a hundred years since Darwin. Explanations have improved since then; we've studied more, we've thought harder. I don't know which is worse sometimes, religious wingnuts who want to poke holes in a theory that no-one credible still stands behind exactly, or "scientific" ignoramii who are still trying to badly defend Darwin exactly as when they learnt it at school.
The religious creationism argument is undebatable on a scientific level anyway; it's not an argument based on rational thought or fact, it's based on ignorance, faith and that human trait of defending the indefensible, simply because it's your indefensible and different from the other guys.
Anyway, I'm a physics geek. I'll put my nukes up against their divine thunderbolts anytime, and we'll see who's a smoking pillar of salt afterwards 8-)
I read more magazines now than I used to 10 years ago. Then I read PC Mag and Byte. Two excellent (at the time) general coverage computer magazines.
Now I read lots of skinny ones, all with a cover price that's twice what Byte used to be. DDJ, VBPJ, MSJ, ObscureNerdWeekly and several others. They're _all_ very niche oriented and written "by nerds, for nerds". Many are subscription only, unless you're within range of about two or three London magazine specialists. As for the newsstand PC comics, I haven't touched one of those in years.
So why aren't I reading them on the Web ? My reading has become narrow-focus specialisation, and isn't that what the Web is best at ? Some of this is still the fact that paper is better than screen (I love my PalmPilot, but it's still not A4 that I can roll up and stuff in a pocket). Much of it is simply that there's still no good way to fund web magazines in a way that allows a publisher to produce something like DDJ. Micro-charging isn't there yet, nor is advertising.
JavaWorld seems to pull it together, but I suspect their revenue model is nothing like that of a traditional magazine. I'd like a good on-line MSJ (M$oft Systems Journal). As it's basically a house technology showcase, I don't know why an on-line version couldn't be produced that did everything paper MSJ does.
Pournelle was originally hired to be their tame idiot, a role for which he is eminently suited.
Byte had obviously lost it when Pournelle started to believe he was one of the most technically literate people left there, and there was no-one else genuinely technical and sufficiently powerful to tell him it wasn't true.
Just look at the expanding / contracting universe theory [...] after this was suggested the scientific community disregarded it and the original author retracted it.
Seems that you don't have much of a clue what "science" means.
Western science doesn't teach facts (as such), it teaches a method. The method (crudely stated) says that a bright idea gets written up, passed around a bit, and described as a "theory". No-one claims it's provably true. It's just there as a hypothetical idea, for discussion and debate. If, after some thought, an experiment is devised that can demonstrate it, then we might start to collect experimental proof that validates it. The best experiment is one that requires some outlandish and unexpected result, but a result that is predicted by this theory. If the experiment then produces that result as predicted, weird as it first sounded, then the majority of scientists start to believe in it.
If after some enormous period of time, a general concensus and a lack of contrary experimental evidence, then the theory may begin to be regarded as a "law of nature". Even then, no-one really claims that it's perfect or entirely accurate; after all Newton's Laws of Motion are demonstrably inaccurate for relativistic speeds, yet we still feel quite happy to build aircraft based on them, nor has anyone suggested they be re-phrased as "Newton's Wrong Theory of Stuff, Hey Isaac, you really like suck, man".
So where does that leave "expanding universe theory" ? Well, it leaves it just there; as a theory. What's your problem here ? No-one ever claimed it was right, just that it was one possible explanation of how things worked, that fitted what was known at the time. We look harder, we think harder, we get better ideas about it. As we've been looking at the universe for barely any time at all, from just the one pipsqueak little planet, then it's amazing we've worked out as much as we have done! Universes are complex critters and they don't come with instruction manuals -- why should we be able to work out how they operate ?
Most of the people who make this comment also suggest a lot of nonsense about adding HTML formatting to texts. This would be a huge mistake.
There's a principle here (quoting from the W3C's WAI) called "The Principle of Closest Markup".
Basically, if you're ever going to do markup, then you should do that markup as soon as possible and as close to the source of the content. It doesn't matter too much what the format is, just that some computer-readable marker gets placed in there while it's still known exactly where the footnotes and paragraph breaks are. It's much easier to re-format from a marked-up text to plaintext than it is to try and automatically add markup to plaintext.
OK, so HTML markup isn't the greatest thing out there, but it has the huge advantage for a project like Gutenberg that everyone (and their dog) knows it. Maybe using a subset like the Slashdot Core would add useful functionality for little cost. If you want plaintext, then stripping those would be trivial scripting.
I think sometimes it hurts you when you stay too long in school I think sometimes it hurts you when you're afraid to be called a fool
Lou Reed / John Cale, "Trouble With Classicists", Songs for Drella
We operate in a chaotic, incompetent industry. Sometimes something is the best innovation ever, sometimes it's a crock that we'll have forgotten in a year. Very rarely can you tell beforehand, and never without actually hands-on experience of it. Analysts operate on the unswerving assumption that everything can be analysed -- any admission that "we can't tell" is a sign of weakness. A "suit"'s greatest failing is that their whole ethos won't allow them to admit any fallibility.
The British solution to this is to let Granny die. It's what we did with the ambulance control fiasco; believe the reassurances of the developers and the suits, and don't do any disaster planning for when it doesn't work.
This is one reason I like the iButton - stronger, less fragile, fewer dead grannies overall.
In this particular context, I don't think that "denial of service" is much of a problem anyway. It's a security- and privacy-critical system, but it's not a time- or life-critical system. After all, this thing isn't treating Granny clinically, just letting her schedule apointments and provide those many slow-turnaround tasks that currently consume a lot of paper and clerical expense in a healthcare situation. What happens if Granny's phone breaks down, or is cut off ? If the WebTeeVee is burgled ?
Burglary is actually a serious point. If WebTeeVees behave anything like VCRs, in a few years they'll be on widespread sale, slighty used, in dodgy pubs. If security is solely on a physical device, which is probably left in the WTV card slot, then stealing someone's screen also steals access to their records. That's going to be a commonplace scenario in the UK, if there isn't a second security identifier.
It's quite possible that appropriate security models are different for the UK and USA. Our healthcare systems are different, and our user demographics are different. In the USA I see this system being adopted so that employed patients can self-schedule appointments to suit working hours and childcare. This is a useful feature, and it would be a selling point for any healthcare plan that offered it. IMHO, it would be a popular feature in the USA.
In the UK, the main consumers of healthcare are in lower social and economic groupings (you have a lousy life, you're going to get ill more). Many can't afford phones, let alone WebTeeVees. Any such system (and they're rumoured regularly) would inevitably be seen as a "healthcare passport" where the primary purpose is to control access to healthcare and limit over-consumption by the undeserving greedy poor (sic). Alternatively, UK liberals (and myself) would see it as a cheapskate government attempting to ration healthcare from the non-voting portion of society. IMHO, this feature could potentially bring the UK to the point of riots.
It stops brute force cracks OK, but it does nothing to stop shoulder surfing and written down passwords. The user community here will contain significant numbers of non computer-literates; they will write the passwords down.
A major security consideration for this system is security of children against parents, and vice versa. Religiously up-tight mother searches daughter's bedroom to find the password on a Post-It and then discovers the birth control prescription ? That's a scenario that's almost guaranteed to turn up sooner or later.
I think it would. Sure, it's an extra level of annoying security - but then do you _really_ want the security problems that plaintext password systems always cause ? It's the designer's call.
Remember that this is a system for patient-doctor communication.
What's a "patient" ? Is this a system that replaces existing phone calls to a receptionist, or does it let the user read their entire medical history, right back to Granny Finkelstein's little "embarassment" back in '68 ? You have to answer this before choosing acceptable security.
NT Challenge / Response isn't a bad way of doing passwords, without throwing plaintext around. No way would I recommend this for a wide-usage public Internet system though - it's just too dependent on client-side installation issues.
Client side certificates work pretty well too. Not easy to deploy to Granny's WTV though.
if you think that they're prepared to ask grandma Finklestein to install a Blue Dot receptor on the web-TV she got for Christmas, you're nuts.
Why not ?
I'm in the UK. If you have anything resembling a webteevee, chances are that it already needs a smartcard inserted to make the cable or satellite link work. Granny's pre-paid GSM mobile phone needs a new SIM inserted into it every month or two, yet the domestic retail market is clearly accepting of this level of technology. This is no longer rocket science.
iButton specifics (I love these things):
use the JavaCard version, unless you're really cutting costs. The fixed ID number buttons can be spoofed by a PIC.
JavaButtons look just like Java SmartCards, if you're writing for them.
JavaButtons aren't anything like Java SmartCards, if you stand on them.
iButton readers work on anything worthy of the name "PC". SmartCard readers are a pain.
Note the fact that this phone is not only limited-life disposable, but also that it's outgoing calls only. This seems obvious at first, after all, who wants a disposable phone with a number that only lasts a month ? Actually it's much smarter than that for a cell phone.
Battery life. An outgoing only phone doesn't need to be left switched on to receive incoming - useful lifetime.
Network traffic. If it can't receive, the network doesn't need to track it. That's a lot less BCH traffic to manage and pay bandwidth for.
Pre-payment charging model. No need for the network operator to bill, no need to log usage duration, no need to verify the phone is valid (unless you're wisely cautious).
No need to negotiate roaming or decent cell handover (I bet it doesn't do this), so that makes the phone simpler.
Batteries can be simple alkaline or lithium primary cells, not awkward rechargeables.
What appears to have been invented here is a disposable version of the old UK Rabbit phone, when in outdoor mode.
Not being rechargable for call time is wasteful, but probably made necessary by limits in reliable crypto. If you ever managed to break the crypto on the call life recharging machine, then you'd have an inexhaustible supply of free calls. No sensible operator would like the idea of being exposed to that risk.
E-speak defines APIs as if they're local procedure calls within the same process (just like our grandparents used to code). There is a wire protocol underlying all that, but you're distanced from it. This is in contrast to XML-RPC / SOAP, where the wire protocol is a defined presentation layer running over a pre-existing session layer (http).
Of course, as it's now Open Source, you can write whatever you need, if you want to, right down to the wire.
This sort of thing is greatly needed,
Indeed. the question is, which one gets taken up by the market. Maybe that just means "Which one does big-cheap-warehouse.com first start offering a price broker service for".
XML-RPC is one, SOAP another, and E-speak yet another.
In a sense, E-speak is (arguably) closer to Corba than to XML-RPC, as it has the same brokering function.
E-speak does something that XML-RPC doesn't. XML-RPC requires clients to communicate with a pre-defined server, from which they have already obtained the metadata that defines their messaging formats. You can't ask an XML-RPC best-price shopping agent to find a new toaster for you, unless you both share a vocabulary that states whether "toaster" is a "product name" or a "product category". If this was easy, then we'd have done it 10 years ago with EDI.
The great advantage of not needing to define the metadata before talking to the shopping agent isn't that it avoids the need to register, but that it means a client that can talk to one about shopping can also talk to any about shopping.
Yes, XML inherently allows self-describing structures based on an extensible syntax that doesn't need metadata to be syntactically validated. They're all good things, but they're still not enough to make it happen.
BizTalk is a little like E-speak, but it brokers the metadata through the wetware at design time, rather than doing it quite so dynamically. Easier to do, but not so flexible in practice.
Information and Content Exchange (ICE) protocol is another contrast. This is a medium-weight protocol that provides a range of resource discovery and trading functions, but gets it to work by restricting the domain of usefulness to that of content syndication (newspapers, weblogs and the like).
I don't speak for HP, and certainly not on E-speak
e"speak isn't open source because HP want collaborative development (in the Mozilla sense), but because of the type of product that it is. e"speak is a protocol, and protocols are only useful if they're adopted by clients. To make it succeed, HP needs to have e"speak adopted by many producers of user agents, browsers and server components and the easiest way to do that is to make the developer's life easy.
It's also no secret that HP has taken a battering from industry analysts lately, and that they publish their quarterly results imminently. e"speak is a deserving product, but the timing is influenced by spin management. I don't speak for HP, BTW.
OTOH, P3P is not a solution to this type of tracking, nor will it ever be. What P3P does is usually misunderstood, even by the nerderati, so please let me point out something significant.
P3P is a protocol for a site to tell a browser what the privacy policy of the site is. Note the direction the information flows in -- only one way. There's no scope in P3P for your browser to be configured to suppress privacy information, nor for it to request a site to not log particular information. The best a full P3P implementation could achieve, even assuming full and honest cooperation of the site operator, is for it to connect to a site and then disable access to the pages with a "Lets not go there" message.
I might still wish to shop at Badgers 'R Us, even though they have a loathsome default logging policy, but only providing they want my business enough to turn logs off on request. Click trails are very low value individually - sites can't afford to lose real trade in favour of them, so we do have the economic advantage here.
P3P can only tell me not to go in, it can't allow me to still shop there without leaving the log trail behind. What we need is a negotiated mechanism for a privacy / logging compromise -- if somewhere like Skylighter (a pyrotechnics vendor) wants to bar users from the shop unless there's reasonable logging in effect, then that's fair and reasonable. OTOH, if World Of Fish request logs, then my browser should tell them to get stuffed and they should either accept this, or lose my business to Piece O' Pike a few blocks down.
A major failing of the UK DPA (Data Protection Act) is that it's too much like P3P. It's good at telling you who has your data, but it's bad at controlling them getting it in the first place. You can't re-bottle the genie.
Slashdot Mirror
This month's DDJ (Dr Dobbs Journal) also has a brief piece on the background to legOS
My son is nearly turned 5
Any /. parents (or loco parents / aunts / uncles / cousins etc.) care to discuss good age ranges for introducing kids to Mindstorms ? It's very easy to hot-house kids into small-parts Lego when they simply don't have the motor skills to assemble the more fiddly parts.
What a pale world you must inhabit, if there's nothing in it more complicated and fascinating than can be explained to a 7 year old in five minutes.
Wilson tubes, and they didn't even need photocells. The "spot" was read as a charge, not by its brightness.
They worked better in the dark, but for debugging you could swing the door over the tube face open and read the dots directly off the screen.
Anyone else remember why "0" has a slash through it ? 8-)
accomodate physically unfit employees
What's "physically fit", and who decides this ?
Ignoring the moral argument for a moment (that unfit or disabled people are just as "deserving" as the fit), then imagine what would happen to the software industry if guys had their code junked because they were too fat ! How many really excellent coders do you know who are (by any reasonable measure) grossly overweight ? Some of the best people around for many jobs are also some of the most unfit, and it's a foolish employer that discards good workers on an irrelevancy.
It is rather unfortunate that the BBC correspondent has very little idea about the subject he is writing about
It's always the same from the Beeb. "Real" intellectuals have arts or humanities degrees, mathematicians are just geeks and beneath contempt. Did you notice the "related links" they placed on the page ? The top feature was last year's "mathematics of biscuit dunking" story. This just shows what little significance the increasingly dumbed-down BBC now places on science or technology stories.
ObRant: Why is it that at the hypothetical mixed-background middle class dinner party, the scientists are expected to be literate, but the literati still revel in their innumeracy ?
It is true that the subject is too esoteric to be accessible to non-mathemticians,
There's probably at least a dozen people in this room as me who work on elliptic curves on a daily basis. OK, so I work in an unusual environment, but these things do actually have real world applications (crypto, natch) and not just for the NSA.
widespread misconception that 1999 is the last year of the 20th century.
So then, no party invites yet ?
Do we care where they hold the insurance?
Yes.
Insurance is a complex business, and prone to many forms of fraud. To prevent this, most countries require that "insurers" comply with some pretty strict (sic) legislation before trading as an insurer. If you're a big company, you can maybe afford to carry your own insurance (and it probably makes sense to do so), but it's likely to be quite illegal to do so unless you also comply with the rules for "being an insurance company".
As an obvious example, consider UPS not delivering something (their fault) and then arguing that it really was delivered (your fault). If they also suffer the financial loss as the insurer, would you really trust them to admit that they'd lost the package and so should pay up ? I'm only a little guy - no way can I (usually) afford to sue huge UPS over it, so being reluctant to pay legitimate claims would make good economic sense for them (not that I claim they'd ever have done this).
Young? But what about old farts like me?
Same as ever. Your working style is as young as your mortgage, partner and kids will permit.
"To live free and happily, one must sacrifice boredom.
It is not always an easy sacrifice."
Richard Bach, I think
(Which I guess qualifies me as an old fart !)
Cray was carrion before Beowulf happened.
The killer for Cray was when large custom boxes full of cheap commodity processors started to appear (Connection Machine et al.). Cray's previous expertise of making hugely clever processors just couldn't compete with Fordian economics and a huge fab churning out Intel games boxes by the bucketload.
Now Beowulf takes it all a step further and replaces a weird box full of standard CPUs with a room full of standard boxes.
Another (quite reasonable) opinion is that Cray was/b> Seymour Cray, and without him they just lost direction.
If there's no "source", how can you be expected to write self-modifying executables ? 8-)
Seriously, I think there's always "source", even for a Lego RCX or Seymour Cray toggling bit patterns into a front panel. Maybe you wrote the executable beforehand, but there has to be something that would pass as a "listing", so that others could replicate it by re-typing what you did.
Darwinian evolution is wrong in big chunks anyway, no-one except a few badly informed scientific bigots still claim it's all right.
It's a hundred years since Darwin. Explanations have improved since then; we've studied more, we've thought harder. I don't know which is worse sometimes, religious wingnuts who want to poke holes in a theory that no-one credible still stands behind exactly, or "scientific" ignoramii who are still trying to badly defend Darwin exactly as when they learnt it at school.
The religious creationism argument is undebatable on a scientific level anyway; it's not an argument based on rational thought or fact, it's based on ignorance, faith and that human trait of defending the indefensible, simply because it's your indefensible and different from the other guys.
Anyway, I'm a physics geek. I'll put my nukes up against their divine thunderbolts anytime, and we'll see who's a smoking pillar of salt afterwards 8-)
I read more magazines now than I used to 10 years ago. Then I read PC Mag and Byte. Two excellent (at the time) general coverage computer magazines.
Now I read lots of skinny ones, all with a cover price that's twice what Byte used to be. DDJ, VBPJ, MSJ, ObscureNerdWeekly and several others. They're _all_ very niche oriented and written "by nerds, for nerds". Many are subscription only, unless you're within range of about two or three London magazine specialists. As for the newsstand PC comics, I haven't touched one of those in years.
So why aren't I reading them on the Web ? My reading has become narrow-focus specialisation, and isn't that what the Web is best at ? Some of this is still the fact that paper is better than screen (I love my PalmPilot, but it's still not A4 that I can roll up and stuff in a pocket). Much of it is simply that there's still no good way to fund web magazines in a way that allows a publisher to produce something like DDJ. Micro-charging isn't there yet, nor is advertising.
JavaWorld seems to pull it together, but I suspect their revenue model is nothing like that of a traditional magazine. I'd like a good on-line MSJ (M$oft Systems Journal). As it's basically a house technology showcase, I don't know why an on-line version couldn't be produced that did everything paper MSJ does.
Pournelle was originally hired to be their tame idiot, a role for which he is eminently suited.
Byte had obviously lost it when Pournelle started to believe he was one of the most technically literate people left there, and there was no-one else genuinely technical and sufficiently powerful to tell him it wasn't true.
Just look at the expanding / contracting universe theory [...] after this was suggested the scientific community disregarded it and the original author retracted it.
Seems that you don't have much of a clue what "science" means.
Western science doesn't teach facts (as such), it teaches a method. The method (crudely stated) says that a bright idea gets written up, passed around a bit, and described as a "theory". No-one claims it's provably true. It's just there as a hypothetical idea, for discussion and debate. If, after some thought, an experiment is devised that can demonstrate it, then we might start to collect experimental proof that validates it. The best experiment is one that requires some outlandish and unexpected result, but a result that is predicted by this theory. If the experiment then produces that result as predicted, weird as it first sounded, then the majority of scientists start to believe in it.
If after some enormous period of time, a general concensus and a lack of contrary experimental evidence, then the theory may begin to be regarded as a "law of nature". Even then, no-one really claims that it's perfect or entirely accurate; after all Newton's Laws of Motion are demonstrably inaccurate for relativistic speeds, yet we still feel quite happy to build aircraft based on them, nor has anyone suggested they be re-phrased as "Newton's Wrong Theory of Stuff, Hey Isaac, you really like suck, man".
So where does that leave "expanding universe theory" ? Well, it leaves it just there; as a theory. What's your problem here ? No-one ever claimed it was right, just that it was one possible explanation of how things worked, that fitted what was known at the time. We look harder, we think harder, we get better ideas about it. As we've been looking at the universe for barely any time at all, from just the one pipsqueak little planet, then it's amazing we've worked out as much as we have done! Universes are complex critters and they don't come with instruction manuals -- why should we be able to work out how they operate ?
Most of the people who make this comment also suggest a lot of nonsense about adding HTML formatting to texts. This would be a huge mistake.
There's a principle here (quoting from the W3C's WAI) called "The Principle of Closest Markup".
Basically, if you're ever going to do markup, then you should do that markup as soon as possible and as close to the source of the content. It doesn't matter too much what the format is, just that some computer-readable marker gets placed in there while it's still known exactly where the footnotes and paragraph breaks are. It's much easier to re-format from a marked-up text to plaintext than it is to try and automatically add markup to plaintext.
OK, so HTML markup isn't the greatest thing out there, but it has the huge advantage for a project like Gutenberg that everyone (and their dog) knows it. Maybe using a subset like the Slashdot Core would add useful functionality for little cost. If you want plaintext, then stripping those would be trivial scripting.
Did anyone else get that ol' reflex twitch and go straight to netnames.org ?
I think sometimes it hurts you when you stay too long in school
I think sometimes it hurts you when you're afraid to be called a fool
Lou Reed / John Cale, "Trouble With Classicists", Songs for Drella
We operate in a chaotic, incompetent industry. Sometimes something is the best innovation ever, sometimes it's a crock that we'll have forgotten in a year. Very rarely can you tell beforehand, and never without actually hands-on experience of it. Analysts operate on the unswerving assumption that everything can be analysed -- any admission that "we can't tell" is a sign of weakness. A "suit"'s greatest failing is that their whole ethos won't allow them to admit any fallibility.
If Gramma's iButton breaks, who she gonna call
The British solution to this is to let Granny die. It's what we did with the ambulance control fiasco; believe the reassurances of the developers and the suits, and don't do any disaster planning for when it doesn't work.
This is one reason I like the iButton - stronger, less fragile, fewer dead grannies overall.
In this particular context, I don't think that "denial of service" is much of a problem anyway. It's a security- and privacy-critical system, but it's not a time- or life-critical system. After all, this thing isn't treating Granny clinically, just letting her schedule apointments and provide those many slow-turnaround tasks that currently consume a lot of paper and clerical expense in a healthcare situation. What happens if Granny's phone breaks down, or is cut off ? If the WebTeeVee is burgled ?
Burglary is actually a serious point. If WebTeeVees behave anything like VCRs, in a few years they'll be on widespread sale, slighty used, in dodgy pubs. If security is solely on a physical device, which is probably left in the WTV card slot, then stealing someone's screen also steals access to their records. That's going to be a commonplace scenario in the UK, if there isn't a second security identifier.
It's quite possible that appropriate security models are different for the UK and USA. Our healthcare systems are different, and our user demographics are different. In the USA I see this system being adopted so that employed patients can self-schedule appointments to suit working hours and childcare. This is a useful feature, and it would be a selling point for any healthcare plan that offered it. IMHO, it would be a popular feature in the USA.
In the UK, the main consumers of healthcare are in lower social and economic groupings (you have a lousy life, you're going to get ill more). Many can't afford phones, let alone WebTeeVees. Any such system (and they're rumoured regularly) would inevitably be seen as a "healthcare passport" where the primary purpose is to control access to healthcare and limit over-consumption by the undeserving greedy poor (sic). Alternatively, UK liberals (and myself) would see it as a cheapskate government attempting to ration healthcare from the non-voting portion of society. IMHO, this feature could potentially bring the UK to the point of riots.
That's very insecure.
It stops brute force cracks OK, but it does nothing to stop shoulder surfing and written down passwords. The user community here will contain significant numbers of non computer-literates; they will write the passwords down.
A major security consideration for this system is security of children against parents, and vice versa. Religiously up-tight mother searches daughter's bedroom to find the password on a Post-It and then discovers the birth control prescription ? That's a scenario that's almost guaranteed to turn up sooner or later.
I don't think that an iButton would work here.
I think it would. Sure, it's an extra level of annoying security - but then do you _really_ want the security problems that plaintext password systems always cause ? It's the designer's call.
Remember that this is a system for patient-doctor communication.
What's a "patient" ? Is this a system that replaces existing phone calls to a receptionist, or does it let the user read their entire medical history, right back to Granny Finkelstein's little "embarassment" back in '68 ? You have to answer this before choosing acceptable security.
NT Challenge / Response isn't a bad way of doing passwords, without throwing plaintext around. No way would I recommend this for a wide-usage public Internet system though - it's just too dependent on client-side installation issues.
Client side certificates work pretty well too. Not easy to deploy to Granny's WTV though.
if you think that they're prepared to ask grandma Finklestein to install a Blue Dot receptor on the web-TV she got for Christmas, you're nuts.
Why not ?
I'm in the UK. If you have anything resembling a webteevee, chances are that it already needs a smartcard inserted to make the cable or satellite link work. Granny's pre-paid GSM mobile phone needs a new SIM inserted into it every month or two, yet the domestic retail market is clearly accepting of this level of technology. This is no longer rocket science.
iButton specifics (I love these things):
This idea isn't as trivial as it first sounds.
Note the fact that this phone is not only limited-life disposable, but also that it's outgoing calls only. This seems obvious at first, after all, who wants a disposable phone with a number that only lasts a month ? Actually it's much smarter than that for a cell phone.
What appears to have been invented here is a disposable version of the old UK Rabbit phone, when in outdoor mode.
Not being rechargable for call time is wasteful, but probably made necessary by limits in reliable crypto. If you ever managed to break the crypto on the call life recharging machine, then you'd have an inexhaustible supply of free calls. No sensible operator would like the idea of being exposed to that risk.
It seems to be a set of APIs
E-speak defines APIs as if they're local procedure calls within the same process (just like our grandparents used to code). There is a wire protocol underlying all that, but you're distanced from it. This is in contrast to XML-RPC / SOAP, where the wire protocol is a defined presentation layer running over a pre-existing session layer (http).
Of course, as it's now Open Source, you can write whatever you need, if you want to, right down to the wire.
This sort of thing is greatly needed,
Indeed. the question is, which one gets taken up by the market. Maybe that just means "Which one does big-cheap-warehouse.com first start offering a price broker service for".
XML-RPC is one, SOAP another, and E-speak yet another.
In a sense, E-speak is (arguably) closer to Corba than to XML-RPC, as it has the same brokering function.
E-speak does something that XML-RPC doesn't. XML-RPC requires clients to communicate with a pre-defined server, from which they have already obtained the metadata that defines their messaging formats. You can't ask an XML-RPC best-price shopping agent to find a new toaster for you, unless you both share a vocabulary that states whether "toaster" is a "product name" or a "product category". If this was easy, then we'd have done it 10 years ago with EDI.
The great advantage of not needing to define the metadata before talking to the shopping agent isn't that it avoids the need to register, but that it means a client that can talk to one about shopping can also talk to any about shopping.
Yes, XML inherently allows self-describing structures based on an extensible syntax that doesn't need metadata to be syntactically validated. They're all good things, but they're still not enough to make it happen.
BizTalk is a little like E-speak, but it brokers the metadata through the wetware at design time, rather than doing it quite so dynamically. Easier to do, but not so flexible in practice.
Information and Content Exchange (ICE) protocol is another contrast. This is a medium-weight protocol that provides a range of resource discovery and trading functions, but gets it to work by restricting the domain of usefulness to that of content syndication (newspapers, weblogs and the like).
I don't speak for HP, and certainly not on E-speak
e"speak isn't open source because HP want collaborative development (in the Mozilla sense), but because of the type of product that it is. e"speak is a protocol, and protocols are only useful if they're adopted by clients. To make it succeed, HP needs to have e"speak adopted by many producers of user agents, browsers and server components and the easiest way to do that is to make the developer's life easy.
It's also no secret that HP has taken a battering from industry analysts lately, and that they publish their quarterly results imminently. e"speak is a deserving product, but the timing is influenced by spin management. I don't speak for HP, BTW.
P3P is a great idea, and I'm all in favour of it.
OTOH, P3P is not a solution to this type of tracking, nor will it ever be. What P3P does is usually misunderstood, even by the nerderati, so please let me point out something significant.
P3P is a protocol for a site to tell a browser what the privacy policy of the site is. Note the direction the information flows in -- only one way. There's no scope in P3P for your browser to be configured to suppress privacy information, nor for it to request a site to not log particular information. The best a full P3P implementation could achieve, even assuming full and honest cooperation of the site operator, is for it to connect to a site and then disable access to the pages with a "Lets not go there" message.
I might still wish to shop at Badgers 'R Us, even though they have a loathsome default logging policy, but only providing they want my business enough to turn logs off on request. Click trails are very low value individually - sites can't afford to lose real trade in favour of them, so we do have the economic advantage here.
P3P can only tell me not to go in, it can't allow me to still shop there without leaving the log trail behind. What we need is a negotiated mechanism for a privacy / logging compromise -- if somewhere like Skylighter (a pyrotechnics vendor) wants to bar users from the shop unless there's reasonable logging in effect, then that's fair and reasonable. OTOH, if World Of Fish request logs, then my browser should tell them to get stuffed and they should either accept this, or lose my business to Piece O' Pike a few blocks down.
A major failing of the UK DPA (Data Protection Act) is that it's too much like P3P. It's good at telling you who has your data, but it's bad at controlling them getting it in the first place. You can't re-bottle the genie.