The original company doesn't have to do this (since, as the author, it doesn't have to release its own changes).
They do if they distribute the binaries. If they're not distributing the binaries, it's hard to see how they can be keeping other people out of the market.
``It happened because of you - users, who wanted to run your current (old) applications on new platform.''
And run them at native speed, too. There's no reason why a 64-bit processor running at 3.0GHz shouldN'T run elderly 8 and 16 bit applications via a SoftPC style instruction-by-instruction emulation, and do so orders of magnitude faster than the original 12MHz (or whatever). But no-one ever grasped the nettle in Microsoft to say ``the performance gap between today's processors and the old ones which need all that glue is such that emulation still gives several orders of magnitude''.
In UK law, administration is closer to US Chapter 11 than Chapter 7; the company is assumed to be capable of being returned to the status of a going concern, rather than a set of assets and liabilities pending liquidation. But in the US the management of the company would normally remain as debtors in possession; in the UK an administrator is more hands-on than a US trustee, has to be agreed by the court, and works for the creditors rather than the shareholders.
ian
You're missing the point. If an adult decides to drop contact with their family and friends, move to another town and start work as a waiter, they have committed precisely no crimes. Firefighters don't comb woods for that: it's a regular event. If there are dependent children then, depending on the law in the country you live in, it's possible you might be financially or criminally liable (oddly, the first parent to leave financially, the second criminally) and if you're married then, depending on your local divorce laws, you might be liable. But if a single man or woman over 18/21/whatever who doesn't have (or continues to pay) debts decides to break contact with their parents and friends, then that's nobody else's business.
As an extension, cases in which couples and their children break contact with one or both sets of parents (ie the children's grandparents) are not uncommon, especially in mixed culture marriages.
They did, however, have Microsoft's blessing. Geeks buy hardware, but everyone else buys systems. The point about Compaq was that they could ship a box that was from an end-user's perspective the same as an IBM-PC, with the active and enthusiastic support of the software vendor who controls said perspective. Compaq's only legal issue was over the BIOS, but as the specification was openly distributed (and was in essence rather similar to CP/M anyway) a cleanroom implementation was hardly a barrier to entry. Microsoft were selling more copies of the software, which they wanted to do, and would support and indeed cheer on Compaq and the rest.
In the case of trying to do the same thing with Apple, at the very best a vendor of clones without Apple's agreement would be able to sail through a very narrow strait on licensing. Apple would have no obligation or enthusiasm to help, and would be legally perfectly OK to erect arbitrary roadblocks in future releases. Arguments that `Microsoft aren't allowed to do that' aren't relevant, because Apple aren't a monopoly: the rules (in most jurisdictions) for monopolists trying to control the market further are rather different.
In the late 1980s I ran X10 and (I think) X11 on Sun 3/50s, with a 15MHz 68020 and 4MB of RAM. It worked well enoug. By the early 90s, X11 was perfectly usable on Sun SLC and ELC machines with 4, 8 or 16MB of RAM. I don't believe that a 1995 33MHz 486 with 8MB of RAM is a substantially slower machine that these workstations of five or ten years previously. Yes, I'm aware that I'm comparing monochrome 1-bit frame buffers with 8-bit colour, but I can't remember the prices, timings or basic performance figures of the Sun IPC which might be a better comparator.
It's not that hard to sell `functionality' in a booming economy, especially when the limitations of the existing product are obvious. But abstruse functionality like WFS, which will provide essentially nothing for end-users because only an idiot is going to market applications which only work on the latest version of Windows, is a pretty hard sell when the economy is heading into recession. Microsoft were able to ride the fact that computers went from ``we all want them to be faster'' to ``yes, that's pretty fast'' during a set of economic cycles. But now computers are ``good enough'', it's going to be a hard sell to get people to upgrade in what is close to a recession. Do I make my mortgage payments or do I buy a machine with an operating system which contains a few filesystem that no applications take advantage of? Difficult choice.
ian
About twenty years I spent a couple of days wandering around the experimental halls at CERN: a friend of my parents was running the data acquisition on an experiment there, and my wife and I had dropped in to see them on our first holiday together. Mike's experiment was looking for quarks (naked bottoms, as I recall), and the events were arriving too frequently for the VAXen they were using. So they had all the sensors rigged to two cables. One short one went off to some custom hardware which made a rough `interesting / not interesting' judgement, and signalled that down another short cable to the VAX. The long one went straight to the VAX, and recorded the signals gated by the output from the hardware device. The difference in the length of the cables was tuned to the settle speed of the hardware devices, so that the gating signal arrived at the same time as the sensor outputs.
Yes, that sort of trick was used by things like Cambridge (slotted) rings. But using propagation delay in your favour always strikes me as really good science thinking.
Your point about Florida is well made, and I suspect that the Carolinas aren't havens of the post-Darwinian synthesis either. Perhaps I should just say ``states that border the Pacific, states that border the Atlantic north of, and including, New York''. Or maybe ``states that contain major research-led Universities''. But I can't help thinking the phrase I'm grasping for is Red and Blue...
think it's safe to say that there's plenty of controversy in the public
In the depressed parts of America, anyway: not unsurprisingly, anywhere in the US within a hundred miles of an ocean is immune from this. But in the fly-over states, large portions of the population are happy to have no health insurance, no jobs, their mortgage foreclosed, their son sent to Iraq and their daughter condemned to an unskilled job by the price of education, just so long as homosexuals a thousand miles away in California can't marry and science textbooks don't have that dang evilution in them.
This is an inland US problem, far less of an issue elsewhere. Creationists are laughable nutters who no-one, no-one takes seriously elsewhere. And as several people have pointed out, neither the Catholic nor the mainstream Protestant churches have any issue with evolution anyway.
``Also, you need to bear in mind that POTS is incredibly simple technology compared to Internet/Cellular/Data services.''
Sat ten feet from a large team attempting to lash POTS support onto a DSLAM, I beg to differ. We all thought it was easy --- hey, we're doing ATM, Ethernet, IP routing, how hard can a bit of baseband voice be? --- but in fact it's damned hard. Consider how you're going to make 999/911/112 work when the switch is melting down under the load of a Pop Idol vote. Hard, eh?
^A and ^E. Handily, every text box that's a product of the standard libraries on a Mac supports (albeit not desperately, or at all, well documented) basic emacs binding. It's a NeXTStep legacy. So I'm typing this into a standard browser text box in Safari, and ^A, ^E, ^T, ^B, ^F, ^K, ^Y, ^O, ^P and ^N have their expected meanings. The meta/escape versions don't work, and there's no marks (^@ or ^-space), kill ring, and so on. But it's enough to be going on with, and makes typing slashdot posts far more civilised.
I would be very upset if someone believed that if both p and q are congruent to 3(mod 4), then p is a perfect square mod q, and q is a perfect square mod p. I'd be upset because it is false. But you know what? I am not going to cry just because no one cares to even understand this.
But there's not a major problem with school boards putting the falsity into state curricula, not are preachers appearing on talk radio demanding that that children not be exposed to such atheistic counterexamples as 15 and 7.
None of the windowing implementations even attempt to be compliant with the GUI guidelines of their host operating system, so they're style breaks. They do weird things with the cut buffers, for example, bashing strange semantics over linking the top of the kill ring with the cut buffer and vice versa. The various windowing packages use the native toolkit to provide something which is neither consistent from platform to platform nor consistent with the ambient look and feel.
They are in some cases hugely invasive into the codebase, and indeed they _do_ clutter up the UI: mouse actions do things which are neither sensible in the usual HCI guidelines of the OS nor sensible in terms of ``it's an xterm'' nor sensible in terms of using the editor (right click and Ctrl+Click are different on 22.1 on OSX, for example, thus breaking usability with a single-button mouse and making a mess for anyone who _knows_ that right clock and ctrl+click are _always_ the same).
It's nice to be able to have several windows in one emacs process (Ctrl-X-5 functions). But the menu and button implementation is horrific.
It's been downhill since the late 17s, in fact. The moment the whole debacle with Epoch, Lucid etc sprung up, emacs became an exercise in people with great programming skills but minimal taste. Windowing support was the last straw: none of them are any good, and they clutter the editor up. Disclaimer: I am not innocent. I wrote the original code that went into the late 17s to provide support for emacs in Suntools. Partly because X11 standardisation was late arriving, emacs got cluttered with Suntools, NeWS, Apollo, X10 and X11 windowing, none of it good enough to be better than leaving the hell alone.
And anyway, although I love GNU emacs to death and I've been using emacsen of various forms for twenty-five years (well, since December 1983), whisper it who dares that actually Greenberg's Multics Emacs had the benefit of being written in genuine MacLisp, including the redisplay loop, whereas GNU Emacs is actually mostly written in C. A trip into the Multics emacs source code is well worthwhile: some of the problems it's solving (redisplay onto vt100 displays down 300 baud circuits) aren't interesting per se, but the approaches most certainly are.
Format and Restore will take care of any password issue if the kid doesn't want to hand it over.
As will a hammer. The idea that one of my kids would provide a computer to another of my kids with the express purpose of bypassing any policies and procedures I have in my own house in anathema. I would also make it quite clear that the elder sibling was not welcome on the premises.
So sibling 1 is providing a computer to sibling 3 that is secured against their common parents, because they don't like what the parents have done to sibling 2's computer? And sibling 3 is seven years old? My, I bet they have fun when they sit down for dinner together.
The idea that it is reasonable to provide for a seven year old a computer to which no responsible adult has access is simply insane. If my nine year old floated that idea to me the MAC address would be barred on the home router in about two seconds, and all access offsite would be transparently proxied into squid as soon as I brought the appropriate instance on air. Anyone who permits a child that young to have unfettered access to the Internet should be sterilised, and anyone who aids and abets them should be treated equally harshly.
Almost all the DSL services you can buy are still over BT's hardware, and BT charge other ISPs by the byte transferred:
It isn't, and they don't. Apart from that, your point is well made. You need to look up LLU, and note the proportion of the market it now covers (indeed, there's half a million lines of LLU being managed from twenty feet from my desk) and then look up how IP Stream and IP Central are priced. Even if an ISP is buying wholesale IP Stream, and many are --- including Vodafone, who are rarely wrong --- and doing their backhaul over IP Central it is not charged by the byte. It is charged by bandwidth: how else would you like it charged?
Of course. Hence my point that this is a great deal more complex than the original poster implied, and has a great deal more opportunities for error. The article was essentially about using fuzzers to force restarts of an AP: if I can kill a router stone dead and force a reboot, the standard urandom mechanism will come up using the same saved state as on the previous boot. I like your idea of using reboot times, but the standard code (and if you can point to a consumer AP manufacturer who is doing security research, please let me know so I can buy their products) only saves one set of state and only does it when the init.d scripts run on the way down. Yes, you can re-write the saved seed out of cron, but standard distributions don't. And doing _that_ has the risk that if I can over power the embedded web server and get it to serve files (a reasonable assumption) I can get the current seed. And so on, and so on.
I don't say any of this is impossible. But it's nothing like as straight forward as ``just generate a random key'', and requires careful study of the risks. WEP is a prime example of how this process goes wrong: the idea wasn't totally unsound, but at every stage minor problems crept in until the reality was utterly useless.
I love it when hands are waved with this degree of enthusiasm. If only it were that easy.
Let look at the problems your ``end of wireless security problem'' has to solve.
You need to prevent a `man in the middle' attack, in which I bring up a rogue base station in the area and have everyone bind to me. Your solution doesn't provide even for a shared secret which I expect the base station to know, so there's nothing to stop this from working. So we're going to need something which a base station can use to prove that it's my base station. What? Certificates? Shared Secrets? All the problems we already have, in fact.
The fine article is mostly about implementation problems, not protocol problems. Both SSH and SSL have been prey to plenty of implementation problems which allow suitably crafted clients to crash, hijack and otherwise mess with servers. You've got all those problems.
And most catastrophically, generating `random keys' in small embedded devices is really, really hard. Getting hold of enough entropy is a small SME router to produce strong keys on a regular basis is difficult. Making sure that initialisation vectors are suitable chosen is hard.
Here's a thought experiment for all `simple' solutions. Imagine I have a router in my lab, the same model as the one I'm attacking. I capture the packets the supplicant sends to initiate an association, and I play them into my captive router. I have the clock on the captive router set an appropriate distance behind the clock of the router I am attacking, and the MAC address set the same and ideally the serial number (they're usually helpfully printed on the outside). What magic is it that makes the key my captive router generates be something other than the key the router I'm attacking generates?
In the UK, a price displayed on an object is an ``offer to treat'', and the vendor has no obligation to sell at that price. There are some exceptions: for example, you can't refuse to sell on the grounds of race, following some ``I won't sell my house to an [epithet]'' cases in the seventies. And there are rules about false advertising which apply, but for which the outcome is _not_ that you get to buy the goods at the advertised price. But in general, if you realise before the point of sale that the pricing is wrong, you don't have to sell. After the sale and you're out of luck, of course.
Now I once bought a £600 CD player for £400 pounds, owing to the small hi-fi shop completely cocking up their inventory, and they phoned me in the evening about it. Once we'd put a small amount of bluster on their side down --- no, sorry, you cannot just ``charge it to my credit card'' and if you try I'll see you in court and win, and how do you fancy running a hi-fi shop with no merchant agreement? --- I had a perfectly amiable conversation with the manager who essentially split the difference with me. I liked the people and didn't want to see the young kid who'd made the mistake pilloried or sacked, and I wanted to be able to shop there again. So I had the £400 player I'd been looking at as well delivered to the house with some CDs, some decent cable and a voucher for a set-up job on my Thorens 160 (this was a long time ago). They realised that I held all the cards, I couldn't be bothered to make enemies, and we parted as best of friends.
ian
Seems like we have a respected scientist confronted with an unexpected phenomena that he wants to understand. What part of that makes him a jackass?
The cold fusion debacle was full of people assuming that F&P really did have heat and neutrons, and then building exotic hypotheses to `explain' this `fact'. Unfortunately F&P didn't have neutrons (or fusion product, or much of anything other than predicted heat from a high=school electrolysis experiment), so the hypotheses were necessarily nonsense. If someone claims an over-unity (or, indeed, too close to unity for comfort) energy transformation, the first task to is figure out if it's real before you try to figure out how it happened. From the quoted article, it appears the academics in question jumped the first step. How do you explain an aluminium prism deflecting N rays? You don't, because it's not real.
Measurement of energy balance in a system is hard. The world of perpetual motion machine cranks is full of experiment that are pure cargo cult.
Interesting question. I don't think it would have eroded that far, but yeah it would have eroded further. But I don't think Vista should be viewed as an end result. Its more a stepping stone. People -needed- to be forced to get off the administrator account for example. That HAD to happen. And that's all by itself has caused a lot of pain.
My only Windows machines at home these days are a few VMs for specific purposes, but even those run with the user logins non-administrators. When there was a Windows machine in production in the house, it similarly had a dedicated admin account and all the users used non-admin accounts. This didn't require a new operating system: it's been possible on NT-series kernels since forever.
people see XP next to OSX
Do they? I doubt many people line the machines up next to each other and go ``hmm, which eye candy do I like best?'' I think the buying decision has been made long before then, especially as more and more Apple get is sold through Apple store and by direct mailorder. I don't know what the factors are that make people choose Apple, but I don't think shiny shiny desktop is that high up the list. Shiny shiny hardware might be: I had surprisingly few problems from my wife for the iMac. I made my original switch to Mac a few years ago on the basis of ``What's the best Unix box I can get that will run Office and therefore shut my management up about the subtle problems with Office-format files that I deliver out of OpenOffice?'' and checking the versions of gcc, perl and emacs that shipped were up to date, but I'm not the mass market.
Slashdot Mirror
ian
``It happened because of you - users, who wanted to run your current (old) applications on new platform.''
And run them at native speed, too. There's no reason why a 64-bit processor running at 3.0GHz shouldN'T run elderly 8 and 16 bit applications via a SoftPC style instruction-by-instruction emulation, and do so orders of magnitude faster than the original 12MHz (or whatever). But no-one ever grasped the nettle in Microsoft to say ``the performance gap between today's processors and the old ones which need all that glue is such that emulation still gives several orders of magnitude''.
ian
In UK law, administration is closer to US Chapter 11 than Chapter 7; the company is assumed to be capable of being returned to the status of a going concern, rather than a set of assets and liabilities pending liquidation. But in the US the management of the company would normally remain as debtors in possession; in the UK an administrator is more hands-on than a US trustee, has to be agreed by the court, and works for the creditors rather than the shareholders. ian
You're missing the point. If an adult decides to drop contact with their family and friends, move to another town and start work as a waiter, they have committed precisely no crimes. Firefighters don't comb woods for that: it's a regular event. If there are dependent children then, depending on the law in the country you live in, it's possible you might be financially or criminally liable (oddly, the first parent to leave financially, the second criminally) and if you're married then, depending on your local divorce laws, you might be liable. But if a single man or woman over 18/21/whatever who doesn't have (or continues to pay) debts decides to break contact with their parents and friends, then that's nobody else's business. As an extension, cases in which couples and their children break contact with one or both sets of parents (ie the children's grandparents) are not uncommon, especially in mixed culture marriages.
http://en.wikipedia.org/wiki/Whole_life_tariff
In the case of trying to do the same thing with Apple, at the very best a vendor of clones without Apple's agreement would be able to sail through a very narrow strait on licensing. Apple would have no obligation or enthusiasm to help, and would be legally perfectly OK to erect arbitrary roadblocks in future releases. Arguments that `Microsoft aren't allowed to do that' aren't relevant, because Apple aren't a monopoly: the rules (in most jurisdictions) for monopolists trying to control the market further are rather different.
ian
In the late 1980s I ran X10 and (I think) X11 on Sun 3/50s, with a 15MHz 68020 and 4MB of RAM. It worked well enoug. By the early 90s, X11 was perfectly usable on Sun SLC and ELC machines with 4, 8 or 16MB of RAM. I don't believe that a 1995 33MHz 486 with 8MB of RAM is a substantially slower machine that these workstations of five or ten years previously. Yes, I'm aware that I'm comparing monochrome 1-bit frame buffers with 8-bit colour, but I can't remember the prices, timings or basic performance figures of the Sun IPC which might be a better comparator.
It's not that hard to sell `functionality' in a booming economy, especially when the limitations of the existing product are obvious. But abstruse functionality like WFS, which will provide essentially nothing for end-users because only an idiot is going to market applications which only work on the latest version of Windows, is a pretty hard sell when the economy is heading into recession. Microsoft were able to ride the fact that computers went from ``we all want them to be faster'' to ``yes, that's pretty fast'' during a set of economic cycles. But now computers are ``good enough'', it's going to be a hard sell to get people to upgrade in what is close to a recession. Do I make my mortgage payments or do I buy a machine with an operating system which contains a few filesystem that no applications take advantage of? Difficult choice. ian
About twenty years I spent a couple of days wandering around the experimental halls at CERN: a friend of my parents was running the data acquisition on an experiment there, and my wife and I had dropped in to see them on our first holiday together. Mike's experiment was looking for quarks (naked bottoms, as I recall), and the events were arriving too frequently for the VAXen they were using. So they had all the sensors rigged to two cables. One short one went off to some custom hardware which made a rough `interesting / not interesting' judgement, and signalled that down another short cable to the VAX. The long one went straight to the VAX, and recorded the signals gated by the output from the hardware device. The difference in the length of the cables was tuned to the settle speed of the hardware devices, so that the gating signal arrived at the same time as the sensor outputs.
Yes, that sort of trick was used by things like Cambridge (slotted) rings. But using propagation delay in your favour always strikes me as really good science thinking.
ian
Your point about Florida is well made, and I suspect that the Carolinas aren't havens of the post-Darwinian synthesis either. Perhaps I should just say ``states that border the Pacific, states that border the Atlantic north of, and including, New York''. Or maybe ``states that contain major research-led Universities''. But I can't help thinking the phrase I'm grasping for is Red and Blue...
This is an inland US problem, far less of an issue elsewhere. Creationists are laughable nutters who no-one, no-one takes seriously elsewhere. And as several people have pointed out, neither the Catholic nor the mainstream Protestant churches have any issue with evolution anyway.
``Also, you need to bear in mind that POTS is incredibly simple technology compared to Internet/Cellular/Data services.'' Sat ten feet from a large team attempting to lash POTS support onto a DSLAM, I beg to differ. We all thought it was easy --- hey, we're doing ATM, Ethernet, IP routing, how hard can a bit of baseband voice be? --- but in fact it's damned hard. Consider how you're going to make 999/911/112 work when the switch is melting down under the load of a Pop Idol vote. Hard, eh?
^A and ^E. Handily, every text box that's a product of the standard libraries on a Mac supports (albeit not desperately, or at all, well documented) basic emacs binding. It's a NeXTStep legacy. So I'm typing this into a standard browser text box in Safari, and ^A, ^E, ^T, ^B, ^F, ^K, ^Y, ^O, ^P and ^N have their expected meanings. The meta/escape versions don't work, and there's no marks (^@ or ^-space), kill ring, and so on. But it's enough to be going on with, and makes typing slashdot posts far more civilised.
None of the windowing implementations even attempt to be compliant with the GUI guidelines of their host operating system, so they're style breaks. They do weird things with the cut buffers, for example, bashing strange semantics over linking the top of the kill ring with the cut buffer and vice versa. The various windowing packages use the native toolkit to provide something which is neither consistent from platform to platform nor consistent with the ambient look and feel. They are in some cases hugely invasive into the codebase, and indeed they _do_ clutter up the UI: mouse actions do things which are neither sensible in the usual HCI guidelines of the OS nor sensible in terms of ``it's an xterm'' nor sensible in terms of using the editor (right click and Ctrl+Click are different on 22.1 on OSX, for example, thus breaking usability with a single-button mouse and making a mess for anyone who _knows_ that right clock and ctrl+click are _always_ the same). It's nice to be able to have several windows in one emacs process (Ctrl-X-5 functions). But the menu and button implementation is horrific.
It's been downhill since the late 17s, in fact. The moment the whole debacle with Epoch, Lucid etc sprung up, emacs became an exercise in people with great programming skills but minimal taste. Windowing support was the last straw: none of them are any good, and they clutter the editor up. Disclaimer: I am not innocent. I wrote the original code that went into the late 17s to provide support for emacs in Suntools. Partly because X11 standardisation was late arriving, emacs got cluttered with Suntools, NeWS, Apollo, X10 and X11 windowing, none of it good enough to be better than leaving the hell alone. And anyway, although I love GNU emacs to death and I've been using emacsen of various forms for twenty-five years (well, since December 1983), whisper it who dares that actually Greenberg's Multics Emacs had the benefit of being written in genuine MacLisp, including the redisplay loop, whereas GNU Emacs is actually mostly written in C. A trip into the Multics emacs source code is well worthwhile: some of the problems it's solving (redisplay onto vt100 displays down 300 baud circuits) aren't interesting per se, but the approaches most certainly are.
The idea that it is reasonable to provide for a seven year old a computer to which no responsible adult has access is simply insane. If my nine year old floated that idea to me the MAC address would be barred on the home router in about two seconds, and all access offsite would be transparently proxied into squid as soon as I brought the appropriate instance on air. Anyone who permits a child that young to have unfettered access to the Internet should be sterilised, and anyone who aids and abets them should be treated equally harshly.
I don't say any of this is impossible. But it's nothing like as straight forward as ``just generate a random key'', and requires careful study of the risks. WEP is a prime example of how this process goes wrong: the idea wasn't totally unsound, but at every stage minor problems crept in until the reality was utterly useless.
ian
- You need to prevent a `man in the middle' attack, in which I bring up a rogue base station in the area and have everyone bind to me. Your solution doesn't provide even for a shared secret which I expect the base station to know, so there's nothing to stop this from working. So we're going to need something which a base station can use to prove that it's my base station. What? Certificates? Shared Secrets? All the problems we already have, in fact.
- The fine article is mostly about implementation problems, not protocol problems. Both SSH and SSL have been prey to plenty of implementation problems which allow suitably crafted clients to crash, hijack and otherwise mess with servers. You've got all those problems.
- And most catastrophically, generating `random keys' in small embedded devices is really, really hard. Getting hold of enough entropy is a small SME router to produce strong keys on a regular basis is difficult. Making sure that initialisation vectors are suitable chosen is hard.
Here's a thought experiment for all `simple' solutions. Imagine I have a router in my lab, the same model as the one I'm attacking. I capture the packets the supplicant sends to initiate an association, and I play them into my captive router. I have the clock on the captive router set an appropriate distance behind the clock of the router I am attacking, and the MAC address set the same and ideally the serial number (they're usually helpfully printed on the outside). What magic is it that makes the key my captive router generates be something other than the key the router I'm attacking generates?ian
In the UK, a price displayed on an object is an ``offer to treat'', and the vendor has no obligation to sell at that price. There are some exceptions: for example, you can't refuse to sell on the grounds of race, following some ``I won't sell my house to an [epithet]'' cases in the seventies. And there are rules about false advertising which apply, but for which the outcome is _not_ that you get to buy the goods at the advertised price. But in general, if you realise before the point of sale that the pricing is wrong, you don't have to sell. After the sale and you're out of luck, of course. Now I once bought a £600 CD player for £400 pounds, owing to the small hi-fi shop completely cocking up their inventory, and they phoned me in the evening about it. Once we'd put a small amount of bluster on their side down --- no, sorry, you cannot just ``charge it to my credit card'' and if you try I'll see you in court and win, and how do you fancy running a hi-fi shop with no merchant agreement? --- I had a perfectly amiable conversation with the manager who essentially split the difference with me. I liked the people and didn't want to see the young kid who'd made the mistake pilloried or sacked, and I wanted to be able to shop there again. So I had the £400 player I'd been looking at as well delivered to the house with some CDs, some decent cable and a voucher for a set-up job on my Thorens 160 (this was a long time ago). They realised that I held all the cards, I couldn't be bothered to make enemies, and we parted as best of friends. ian
ian
That story is often told (with fifty years instead of five hundred years) of Aberdeen...