Doesn't want to be faced with large numbers of complaints about the action, some at least of which will be ill-considered and ill-mannered,
Doesn't want to court/ risk bad publicity,
Doesn't want Suse to receive large numbers of expressions of support, some at least of which will be well-considered and may help Suse resist the claim.
Or something else entirely. Until more details are known it's too soon to say.
How powerful would one need to be, I wonder? And could you keep a good enough watch to have enough time to build a custom asteroid-killer when one becomes needed?
Place I work used "happymail" for years as its primary backbone. Used another package for calendaring. Then HP started emphasising integration with Windows clients and offloading non-core connector modules onto third party suppliers, and the calendaring integration (as far as I've heard) lagged behind: the overal impression was that HP was less than fully committed to the product. So Exchange got the nod for the future. And once that sort of decision gets taken high up in a large company, it's seldom turned round - amongst other reasons because some high-ups would have to admit they'd made a mistake. In fact, from their point of view this takeup by Samsung supports their decision: the original vendor is backing out of the product, why should they continue to depend on it?
Another poster mentioned Dec's MailWorks and its descendant: there yoiu have another case of an excellent product (for its time) that sank because of lack of commitment by the vendor. (Well, you can say that about Digital in general near the end, but MailWorks was one of the earlier signs.)
Sorry, folks, but at the level where OM was pitched, perceptions are very important.
"The default hard disk icon is part of the desktop metaphor."
Funny, I always thought it was complementary to the desktop metaphor.
If you're looking for ease of use for a limited set of functions, by all means put icons on the desktop, or group then into function-related folders on the desktop, or whatever. Have more than one desktop each with its own set of icons or folders for mutually exclusive functions? By all means. But do provide a reasonable way for system managers to easily organise the functions in a way that makes sense from the user's point of view. To some extent, this is already done when you (as end-user and system manager of your own personal Wintel box) install an intelligently-packaged new application and are asked whether you want an icon for it on the desktop, and where you want it to be integrated into the taskbar mechanism. Of course, some application vendors believe that their customers shouldn't have these choices, but that's another matter.
But once you get beyond a certain number of functions, and a certain level of complexity, then direct access to the underlying hierarchical file system has a lot to recommend it.
One thing I would want to know before entrusting any significant identifying information to any of these proposed services:
How will the services be audited, under what criteria, by whom, and with what penalties for failing to meet the criteria?
And yes, I do think there should be a law about it - whether it's the application of an appropriate existing one or (shudder) a new one that takes account of the possibility - or should one say the service providers' intention - that the service will be used for a wide range of matters which have until now not been related or connected in any way.
I challenge anyone, anywhere, to give evidence that anyone has ever collected a single penny from suing a mass-market software maker for shoddy code.
mass-market != enterprise
Much of the work that goes into deploying enterprise-level solutions is concerned precisely with ensuring that there is accountability, so that when problems occur it is clear whose responsibility it is to resolve them, and how quickly.
The "who can we sue" argument is nonsensical to the extent that if you get to that point, you're already bleeding resources or at worse going out of business. But it's a great comfort to know that if the project goes bad there'll be a way to muddy the waters about who was to blame for a bad choice at the initial planning stage.
The bit I don't like the sound of is (after discussion of another company's policy)
At Dollar, the rental agreement forms -- and thumbprints -- are stored at the company's corporate headquarters in Tulsa, Oklahoma, for seven years before they are trashed.
I wouldn't have any particular problem with giving a thumbprint provided it's retained only as long as is reasonable to cover theft of the vehicle and fraudulent use of credit card or other identity.
But seven years seems an awful long time to keep records of this sort. One hopes the records repository and the forwarding of the records are all secure....
I also wonder how long the company expects to keep a competitive advantage by doing this - one assumes that it will discourage at least some of the criminal fraternity from going there as a first choice, but if it works that well, other companies may well take it up.
But frankly, I'd go further. You've got a project that's complicated enough to have an 18-month schedule for design through to initial deployment and real-life testing, but which is then to be migrated to a significantly different system platform after only a furhter 12 months. Now, there may be sound reasons for this which people not familiar with the circumstances can't know, but to me it sounds distinctly odd, and frankly not the sort of project I'd want to get involved in. I don't suppose it would be possible to get back to the client and (tactfully) point out that moving to a different platform is going to cost a significant fraction of the initial implementation? (Off the top of my head I'd reckon at least 50%, because all the validation and quality assurance is going to have to be done all over again, even assuming fullest modularity in the original work.) Perhaps, just perhaps, starting on the intended longterm platform will then look more attractive.
Or, perhaps they can consider staying on the initial platform for longer than currently envisaged (which may be what happens anyway once the work involved in migrating becomes clear). I'm more comfortable myself working under Unix, but NT on good-quality hardware running well within its capacity and performance limits is a viable platform for well-defined application workloads. We have some specialised machines where I work that typically run 3 months or so between precautionary reboots, and have run >6 months when operations forgot on one occasion.
Unfortunately, I think Rogerborg's assessment is a fair take on this settlement. Microsoft will probably take care not to test the envelope of the specifically-enjoined limitations for a while, at least not in ways in ways that will be blatantly obvious. But the company's history and its behaviour throughout the period of the trial indicate that it will consider that it is free to invent other ways of consolidating and controlling existing and new markets for as long as these are not explicitly forbidden.
Judge Jackson ordered the company to be split up because, basically, he had come to the conclusion that its senior officers could not be trusted to stick to promises and undertakings they entered into. The appeals court threw out this remedy partly on the grounds that Judge Jackson had allowed his personal dislike for the attitudes of senior MS officials including Mr Gates to affect his decision. Arguably, this was missing the point by several light-years: Microsoft's corporate culture is - at least - hugely influenced by Mr Gates' attitudes and beliefs. It is difficult, for example, to imagine that Mr Balmer would have performed that amazing motivational monkey-dance if Mr Gates was known to be thinking Deep Thoughts about whether the business practices that were appropriate for a small startup company challenging the industry giants were still appropriate for, well, an adjudged monopolist.
Perhaps Mr Gates will be thinking about this in future. The settlement does at least (IV.C) require that the company appoints a Compliance Officer whose duties include briefing senior officers annually about the settlement and the antitrust laws in general (IV.C.3.c), and collecting their signed assents that they understand and will abide by the terms of the settlement (IV.C.3.d - interesting there's no mention of antitrust laws in general in this clause, another example of wriggle-room, perhaps?).
But it's disappointing that the Compliance Officer's mandated activities don't include initiating and running an educational campaign so that all MS staff are informed of the nature and reasons for anti-trust legislation, the reasons that the company was found to be in breach of them, and the general restraints that the legislation imposes on the activities of companies that have significant share of their markets.
To take an example, I work in IT in the financial sector, without any direct contact to my employer's clients. Nevertheless, my work contract contains clauses covering both customer confidentiality and about my obligation to report any financial activities that raise questions about fraud, money-laundering, etc. And I'm regularly reminded of these things by notices on the company intranet, messages passed through line management, and from time to time by printed memos from the chief exec to all staff. Now, if Mr Gates were to head up a similar drive to acquaint all staff with Microsoft's obligations under antitrust, then I might believe that his company was genuinely changing....
That [a license fee] is how the BBC is supported IIRC.
In this case, "license" = mispelling of "tax". The only condition on being issued with this type of license is that you pay the money demanded for it. Contrast driving license, license to practice medicine, poetic license, etc, for which you must demonstrate that you're capable of doing whatever is involved.
No matter what, after Sept. 11, there will be some serious security measures on airports and other problematic zones. These national ID cards are actually a convenient way how to avoid these.
Although I know this isn't what you meant, I'm afraid that "convenient way to avoid [serious security measures]" is exactly why ID cards are so tempting to the authorities: the promise of getting something on the cheap. If you really want air travel to be reasonably secure from hijacking etc then you're going to have to put up with the inconvenience and added expense of doing the job properly. For example, El Al's procedures which start with check-in times long enough to do serious checks on luggage and on passengers' reasons for travelling, and security staff who are paid above minimum wage.
My feeling is that ID cards become more effective the less they are asked to do. Like many other people, I have an ID badge that secures my entry to the place I work. Doubtless other companies use the same technology, but with their own encoded data on the cards, so stealing one from company X isn't going to help you much getting into company Y's premises. And though I can get into the office area of my main work location without it if I carelessly leave the badge at home, the procedures to do so are deliberately made to be inconvenient to discourage laxness.
By contrast, a single card that authenticates you for a whole range of purposes which you might well not have need for in the immediate future sounds a much more attrative target for theft.
Presumably, there would have to be some method of identifying a legally encrypted message from an illegally encrypted one.
Oh, that's trivial. The essence of the UK's pernicious RIP Act is that it's illegally encrypted if the government - or any of the myriad official agencies that the responsible minister can choose to designate as and when he feels like it - cannot decrypt it and you cannot or will not decrypt it for them on demand (modulo persuading the authorities/ a court that you cannot reasonably be expected to be able to - but the burden of such proof is on you, not on the authorities to demonstrate the opposite "beyond reasonable doubt").
Go to jail. Go directly to jail. Do not pass Go. Do not collect two hundred pounds.
Easy enough to quickly turn any politician who votes for such a stinker into a criminal.
Er, no. This possibility was demonstrated by opponents to the UK's Regulation of Investigatory Powers (RIP) bill by sending just such an encrypted message to the then Home Secretary (Interior Minister), now Foreign Secretary Jack Straw, when the legislation was at an early stage in its path through the UK parliament.
The result? The RIP bill's language was adjusted so that ministers of the crown and other duly authorised people would be exempted from the consequences of such underhanded malicious tricks. No help for the rest of us, we would still have to persuade a court of the plausibility of our contentions that we don't know the relevant key.
I was also thinking a replicated database may be a partial answer, but not at airports.
Centralised complete database, use limited to high-sensitivity locations which need access to the complete information in real time. For that matter, where the application is just to authenticate an indentity, you can split up the data into multiple backend engines and have a switch on the front to route the request to the correct one based on the id that's being checked.
There's more than one way to scale or distribute a thing like this.
In one respect, on a matter that's seen some discussion, it doesn't matter a great deal whether strong encryption is being used or not. If you rely overmuch on electronic eavesdropping and code-breaking, then you're in trouble if the black-hats use effectively unbreakable or untraceable hightech, and if they use effectively untraceable lowtech.
Either way, what you're going to need is to get your hands dirty by the traditional methods - infiltration into the blackhat organisations or at least into areas sympathetic enough to hear interesting rumours.
I agree with the consensus: changing the Legos name is a trivial price to pay for maintaining the company's intelligent and friendly attitude to genuine innovation. Pity that any replacement name is unlikely to be as appropriate, succinct and descriptive, but LogOS achieves that precisely because it includes the trademark. So, has anyone got some ideas for neat new names? Best I've come up with so far are
Mindstorms(r) Dynamic Operating System - unfortunately the acronym clashes with a trademark held by a company with a more selective interpretation of the term innovation.
Someone made a joke about Gnu/LegOS, but along the same lines, and more seriously:
GLOSS - Gnu Lego(r)/Mindstorms(r) Open Source System? But dragging the O-word and its associated political baggage into what has so far been a friendly relationship between the brick-hackers and the brick-vendor is probably not a Good Idea. So,
GLASS - Generic lego(r)/Mindstorms(r) Alternative Software System? Hint of transparent access to the inards of the brick, and that it's not official.
I suspect the bit about sympathetic vibrations causing a bridge to collapse is mostly an urban legend, at least for modern bridges.
Tell that to Arup, Foster, and Caro. Although it didn't collapse, the so-called Millenium Bridge which they were responsible for over the Thames in London had to be closed to the public a day or so after being declared open because it swayed too much. That was in the middle of last year, and as far as I know it's still closed. Apparently one of its resonances was near enough to a multiple of average walking pace to cause trouble, and once the swaying started, people's reactions tended to syncronise their movements to the bridge's. They're talking about fitting expensive dampers - but perhaps a few retired army officers would do as well:
"Break step, you horrible little people! Break step, I say!"
For those/.ers who are interested enough about DMCA to want to write to their representatives in Congress about it, there is a note in the latest Risks Digest from another person who has concluded that he cannot take the risk of continuing to work in the encryption field either commercially or academically. Tracing the back links in that story gives other instances.
... Wiretap warrants require much more procedure than simple search warrants, and are mainly a creation of telephone regulatory law that's not clearly applicable here, since the Consitutional justification for telephone wiretaps is that the phone company is outside your house....
So, can a non-IANAL clarify what authorisation is required for investigatory authorities to install bugging devices of any type in a suspect's private property?
Any guesses what they were doing wrong with the "massive combat simulation study"?
To state the obvious, wrong assumptions that could only be properly tested in real combat. Don't forget that that there'd never been anything that heavy, fast, and long-range before, and that the development program had to be pushed through at an incredible tempo once the war started in the Pacific where such an aircraft was essential. It's not surprising that simulations didn't always get the right answers - what was critical was that the real-world results were accepted and acted on, rather than a having a lot of arguing with the simulations people fighting for "their" results.
There's a lot of material available online about the plane. Try asking Google about "B-29" and combat and warbirds, and look at the warbirdsresourcegroup page near the top of the list.
One other point I overlooked. This is a city government operation, which I would bet has less need to interchange "industry-standard" formatted files with other organisations than the typical commercial company, so it can better tolerate being "non-standard".
This is a story about an organisation that's been using thin client devices (what we used to call X-terminals, and NCD have been making them for quite a few years) connected to backed servers. Moreover, the organisation has been using this model since 1992, with various backend server platforms. Currently, they're using Linux as the backend, and it's quite likely a good deal more attractive and easier to administer than the Unixware that preceded it, but remember that they've now had (assuming at least 2 sysmans throughout that time, probably more when they were initially starting with it) over 18 man-years of tuning and replacing components with better technology as it became available. Don't get me wrong, this sounds an impressive achievement, but if the organisation were to put out a request for tender for a new IT system today, would it take the risk of going with such an unusual approach?
I think we all know the answer to that: they'd go with industry-standard desktop PCs with industry-standard software. Because it would be "less risky".
Perceptions are important
on
Windows in 2020
·
· Score: 2, Interesting
I find it interesting that articles like this - whether written humourously or seriously (eg this one from the current Economist) are now no longer rare in the mainstream press. Distrust of Microsoft in general and its aggressive business practices in particular are no longer restricted to the IT-knowledgable.
Whether this distrust will result in the company being constrained to operate consistently with its monopoly status is - unfortunately - another matter entirely. It looks as though MS's top management has decided to construct new facts on the ground which will make current court rulings irrelevant; unfortunately the US justice system appears unable to cope with this strategy.
A high level of interrupts is bad for throughput, too.
But the tendency to use words and analogies drawn from current technology has a long history. Popular-science accounts of the working of the brain used to compare it with a telephone exchange. At the time they were written, this was the highest vaguely relevant technology. Fifty years later, comparisons were being made with computers. Most such analogies and comparisons become misleading if you try to extrapolate from the analogy back to real life - the brain isn't a telephone exchange or a computer, after all.
Vaguely related: long ago, when the organisation I was in had far too much work for its headcount (something to do with a recession and layoffs - little changes in business) our group manager once picked on 'concurrency' as his word of the week. "Up your concurrency!" he exhorted his staff at an open meeting. A prim female voice from the back of the hall responded "Up yours."
Slashdot Mirror
- Doesn't want to be faced with large numbers of complaints about the action, some at least of which will be ill-considered and ill-mannered,
- Doesn't want to court/ risk bad publicity,
- Doesn't want Suse to receive large numbers of expressions of support, some at least of which will be well-considered and may help Suse resist the claim.
Or something else entirely. Until more details are known it's too soon to say.Anyone seen any discussions of this?
Another poster mentioned Dec's MailWorks and its descendant: there yoiu have another case of an excellent product (for its time) that sank because of lack of commitment by the vendor. (Well, you can say that about Digital in general near the end, but MailWorks was one of the earlier signs.)
Sorry, folks, but at the level where OM was pitched, perceptions are very important.
Funny, I always thought it was complementary to the desktop metaphor.
If you're looking for ease of use for a limited set of functions, by all means put icons on the desktop, or group then into function-related folders on the desktop, or whatever. Have more than one desktop each with its own set of icons or folders for mutually exclusive functions? By all means. But do provide a reasonable way for system managers to easily organise the functions in a way that makes sense from the user's point of view. To some extent, this is already done when you (as end-user and system manager of your own personal Wintel box) install an intelligently-packaged new application and are asked whether you want an icon for it on the desktop, and where you want it to be integrated into the taskbar mechanism. Of course, some application vendors believe that their customers shouldn't have these choices, but that's another matter.
But once you get beyond a certain number of functions, and a certain level of complexity, then direct access to the underlying hierarchical file system has a lot to recommend it.
Just my 0.02 Euros
How will the services be audited, under what criteria, by whom, and with what penalties for failing to meet the criteria?
And yes, I do think there should be a law about it - whether it's the application of an appropriate existing one or (shudder) a new one that takes account of the possibility - or should one say the service providers' intention - that the service will be used for a wide range of matters which have until now not been related or connected in any way.
mass-market != enterprise
Much of the work that goes into deploying enterprise-level solutions is concerned precisely with ensuring that there is accountability, so that when problems occur it is clear whose responsibility it is to resolve them, and how quickly.
The "who can we sue" argument is nonsensical to the extent that if you get to that point, you're already bleeding resources or at worse going out of business. But it's a great comfort to know that if the project goes bad there'll be a way to muddy the waters about who was to blame for a bad choice at the initial planning stage.
The bit I don't like the sound of is (after discussion of another company's policy)
I wouldn't have any particular problem with giving a thumbprint provided it's retained only as long as is reasonable to cover theft of the vehicle and fraudulent use of credit card or other identity.But seven years seems an awful long time to keep records of this sort. One hopes the records repository and the forwarding of the records are all secure....
I also wonder how long the company expects to keep a competitive advantage by doing this - one assumes that it will discourage at least some of the criminal fraternity from going there as a first choice, but if it works that well, other companies may well take it up.
Al
But frankly, I'd go further. You've got a project that's complicated enough to have an 18-month schedule for design through to initial deployment and real-life testing, but which is then to be migrated to a significantly different system platform after only a furhter 12 months. Now, there may be sound reasons for this which people not familiar with the circumstances can't know, but to me it sounds distinctly odd, and frankly not the sort of project I'd want to get involved in. I don't suppose it would be possible to get back to the client and (tactfully) point out that moving to a different platform is going to cost a significant fraction of the initial implementation? (Off the top of my head I'd reckon at least 50%, because all the validation and quality assurance is going to have to be done all over again, even assuming fullest modularity in the original work.) Perhaps, just perhaps, starting on the intended longterm platform will then look more attractive.
Or, perhaps they can consider staying on the initial platform for longer than currently envisaged (which may be what happens anyway once the work involved in migrating becomes clear). I'm more comfortable myself working under Unix, but NT on good-quality hardware running well within its capacity and performance limits is a viable platform for well-defined application workloads. We have some specialised machines where I work that typically run 3 months or so between precautionary reboots, and have run >6 months when operations forgot on one occasion.
Good luck.
Good questions, and answers overflowing with goodwill and scholarship. Excellent work, all round.
Judge Jackson ordered the company to be split up because, basically, he had come to the conclusion that its senior officers could not be trusted to stick to promises and undertakings they entered into. The appeals court threw out this remedy partly on the grounds that Judge Jackson had allowed his personal dislike for the attitudes of senior MS officials including Mr Gates to affect his decision. Arguably, this was missing the point by several light-years: Microsoft's corporate culture is - at least - hugely influenced by Mr Gates' attitudes and beliefs. It is difficult, for example, to imagine that Mr Balmer would have performed that amazing motivational monkey-dance if Mr Gates was known to be thinking Deep Thoughts about whether the business practices that were appropriate for a small startup company challenging the industry giants were still appropriate for, well, an adjudged monopolist.
Perhaps Mr Gates will be thinking about this in future. The settlement does at least (IV.C) require that the company appoints a Compliance Officer whose duties include briefing senior officers annually about the settlement and the antitrust laws in general (IV.C.3.c), and collecting their signed assents that they understand and will abide by the terms of the settlement (IV.C.3.d - interesting there's no mention of antitrust laws in general in this clause, another example of wriggle-room, perhaps?).
But it's disappointing that the Compliance Officer's mandated activities don't include initiating and running an educational campaign so that all MS staff are informed of the nature and reasons for anti-trust legislation, the reasons that the company was found to be in breach of them, and the general restraints that the legislation imposes on the activities of companies that have significant share of their markets.
To take an example, I work in IT in the financial sector, without any direct contact to my employer's clients. Nevertheless, my work contract contains clauses covering both customer confidentiality and about my obligation to report any financial activities that raise questions about fraud, money-laundering, etc. And I'm regularly reminded of these things by notices on the company intranet, messages passed through line management, and from time to time by printed memos from the chief exec to all staff. Now, if Mr Gates were to head up a similar drive to acquaint all staff with Microsoft's obligations under antitrust, then I might believe that his company was genuinely changing....
(OK, I lied about the poetic license. So sue me.)
Although I know this isn't what you meant, I'm afraid that "convenient way to avoid [serious security measures]" is exactly why ID cards are so tempting to the authorities: the promise of getting something on the cheap. If you really want air travel to be reasonably secure from hijacking etc then you're going to have to put up with the inconvenience and added expense of doing the job properly. For example, El Al's procedures which start with check-in times long enough to do serious checks on luggage and on passengers' reasons for travelling, and security staff who are paid above minimum wage.
My feeling is that ID cards become more effective the less they are asked to do. Like many other people, I have an ID badge that secures my entry to the place I work. Doubtless other companies use the same technology, but with their own encoded data on the cards, so stealing one from company X isn't going to help you much getting into company Y's premises. And though I can get into the office area of my main work location without it if I carelessly leave the badge at home, the procedures to do so are deliberately made to be inconvenient to discourage laxness.
By contrast, a single card that authenticates you for a whole range of purposes which you might well not have need for in the immediate future sounds a much more attrative target for theft.
Go to jail. Go directly to jail. Do not pass Go. Do not collect two hundred pounds.
The result? The RIP bill's language was adjusted so that ministers of the crown and other duly authorised people would be exempted from the consequences of such underhanded malicious tricks. No help for the rest of us, we would still have to persuade a court of the plausibility of our contentions that we don't know the relevant key.
Centralised complete database, use limited to high-sensitivity locations which need access to the complete information in real time. For that matter, where the application is just to authenticate an indentity, you can split up the data into multiple backend engines and have a switch on the front to route the request to the correct one based on the id that's being checked.
There's more than one way to scale or distribute a thing like this.
Either way, what you're going to need is to get your hands dirty by the traditional methods - infiltration into the blackhat organisations or at least into areas sympathetic enough to hear interesting rumours.
Mindstorms(r) Dynamic Operating System - unfortunately the acronym clashes with a trademark held by a company with a more selective interpretation of the term innovation.
Someone made a joke about Gnu/LegOS, but along the same lines, and more seriously:
GLOSS - Gnu Lego(r)/Mindstorms(r) Open Source System? But dragging the O-word and its associated political baggage into what has so far been a friendly relationship between the brick-hackers and the brick-vendor is probably not a Good Idea. So,
GLASS - Generic lego(r)/Mindstorms(r) Alternative Software System? Hint of transparent access to the inards of the brick, and that it's not official.
I'm sure others can do better.
Tell that to Arup, Foster, and Caro. Although it didn't collapse, the so-called Millenium Bridge which they were responsible for over the Thames in London had to be closed to the public a day or so after being declared open because it swayed too much. That was in the middle of last year, and as far as I know it's still closed. Apparently one of its resonances was near enough to a multiple of average walking pace to cause trouble, and once the swaying started, people's reactions tended to syncronise their movements to the bridge's. They're talking about fitting expensive dampers - but perhaps a few retired army officers would do as well:
"Break step, you horrible little people! Break step, I say!"
There's more information about it all here.
For those /.ers who are interested enough about DMCA to want to write to their representatives in Congress about it, there is a note in the latest Risks Digest from another person who has concluded that he cannot take the risk of continuing to work in the encryption field either commercially or academically. Tracing the back links in that story gives other instances.
So, can a non-IANAL clarify what authorisation is required for investigatory authorities to install bugging devices of any type in a suspect's private property?
To state the obvious, wrong assumptions that could only be properly tested in real combat. Don't forget that that there'd never been anything that heavy, fast, and long-range before, and that the development program had to be pushed through at an incredible tempo once the war started in the Pacific where such an aircraft was essential. It's not surprising that simulations didn't always get the right answers - what was critical was that the real-world results were accepted and acted on, rather than a having a lot of arguing with the simulations people fighting for "their" results.
There's a lot of material available online about the plane. Try asking Google about "B-29" and combat and warbirds, and look at the warbirdsresourcegroup page near the top of the list.
One other point I overlooked. This is a city government operation, which I would bet has less need to interchange "industry-standard" formatted files with other organisations than the typical commercial company, so it can better tolerate being "non-standard".
This is a story about an organisation that's been using thin client devices (what we used to call X-terminals, and NCD have been making them for quite a few years) connected to backed servers. Moreover, the organisation has been using this model since 1992, with various backend server platforms. Currently, they're using Linux as the backend, and it's quite likely a good deal more attractive and easier to administer than the Unixware that preceded it, but remember that they've now had (assuming at least 2 sysmans throughout that time, probably more when they were initially starting with it) over 18 man-years of tuning and replacing components with better technology as it became available. Don't get me wrong, this sounds an impressive achievement, but if the organisation were to put out a request for tender for a new IT system today, would it take the risk of going with such an unusual approach?
I think we all know the answer to that: they'd go with industry-standard desktop PCs with industry-standard software. Because it would be "less risky".
Whether this distrust will result in the company being constrained to operate consistently with its monopoly status is - unfortunately - another matter entirely. It looks as though MS's top management has decided to construct new facts on the ground which will make current court rulings irrelevant; unfortunately the US justice system appears unable to cope with this strategy.
A high level of interrupts is bad for throughput, too.
But the tendency to use words and analogies drawn from current technology has a long history. Popular-science accounts of the working of the brain used to compare it with a telephone exchange. At the time they were written, this was the highest vaguely relevant technology. Fifty years later, comparisons were being made with computers. Most such analogies and comparisons become misleading if you try to extrapolate from the analogy back to real life - the brain isn't a telephone exchange or a computer, after all.
Vaguely related: long ago, when the organisation I was in had far too much work for its headcount (something to do with a recession and layoffs - little changes in business) our group manager once picked on 'concurrency' as his word of the week. "Up your concurrency!" he exhorted his staff at an open meeting. A prim female voice from the back of the hall responded "Up yours."