Because one day Microsoft will ban you from using anything that resembles what you now call "Windows". In the interest of providing you a better service, of course, and nothing to do with trying to get more revenue. Or Microsoft will die. Computer companies are short-lived things, compared to their customers.
The patch they released 21 days ago was the one that broke SHARE detection so that all our users were getting "Missing SHARE.EXE" reports from their applications. We had to patch our app to work with the patched XP. Fortunately users are so used to Microsoft buggering things up that they accept it as a perfectly normal part of life and don't claim that we're lying to get ourselves out of trouble.
Pass legislation enabling anyone who buys a product as a result of receiving a spam to repudiate the charge on his credit card. The card companies will start vetting their merchants more thoroughly PDQ. Money=muscle.
All that is needed is the enactment of a simple law: if you purchase any product as a result of receiving spam, you are free to repudiate the debt. In particular, if your credit card was charged, your credit card company must credit your account on receiving evidence that you made the purchase in response to a spam. This avoids all the tedious process of law enforcement: no-one gets prosecuted, spam remains entirely legal, free speech isn't affected...
We let our users password-protect their databases. So of course they lose the passwords and we have to have a mechanism (challenge-response) to let them break in and reset the master password. The problem is, how do we know the person asking for this service is the owner of the data? There's no way (that I can see) of both guaranteeing that a thief won't ask to have his password broken into and that a legitimate owner won't be prevented from rescuing his own data.
Some time in the last century (shortly after PCs had been invented), we created an estate agency software package called Monopoly. PC User published a review of it and illustrated it with a photograph of a certain property trading board game.
We got a letter from Waddingtons' lawyers. We wrote back to them saying that (a) we were not using "Monopoly" as the name of a game; (b) we were not implying endorsement by Waddingtons or association with their game; (c) we were not mentioning their game in any way in our documentation or marketing material.
They didn't have a leg to stand on and they didn't even reply. We carried on using the name.
If it's a registered trademark then it is registered for particular classes of goods. For example, VAX is the trademark of a vacuum cleaner, of a type of computer, and of coloured felt-tips. Three different trademark classes, three different owners, no conflict. You must not use the trademark to denote goods in those classes. In addition, you must not imply association with the trademark or promote confusion among the public. If you are using "hard radio" to mean something that is nothing to do with what HardRadio is or does, (eg. if they are using it to make people think of "hard-core radio" and you are using "hard radio" to mean radio implemented by hardware rather than software) then they don't have much of a case. If you are using it to promote something similar with or competitive to them, then they have a very good case. Remember what trademarks are for, and you won't go far wrong.
This is UK rather than US law. As far as I can see, it would be quite reasonable under UK law for someone to market a brand of too-small too-tight condoms under the name Microsoft... as long as they could show that they were using the brandname to sell the condoms and not merely to devalue the trademark of another company. Even if the case went to court, it would hinge on whether such devaluation was actually likely.
Let's go for Venus instead. Lots of raw materials to play with, and it'll be fun to design micro-organisms that have a lifecycle floating in the clouds. The bacteria from terrestrial black smokers will be really surprised to find themselves in their new location....
You're probably covered for $2-3m already, under one policy or other, if (say) you are careless crossing the road and cause a major accident. PC risks are just as insurable - and it will be interesting to see how the insurance market prices $1m of cover for Windows systems versus Linux systems versus...
You mean, you can get the cashier at Safeway to shout out loud any word you want? Don't complain, enjoy!
Mr Ihoputa - it's a real Japanese surname. Ms Foyame - especially if the cashier is female and cute. [there's an acute accent on the final E]. Mr Chupapoyas - especially if the cashier is male and camp.
This applies to the southern USA, but Canadians can have their own equivalent.
There is no such thing as "the Linux GUI" and until such a thing exists no serious porting effort of commercial software takes place. I don't care which **the** Linux GUI is but I am not going to (a) run two parallel development projects or (b) bet my company on whether VHS or Beta will win in the marketplace. It doesn't matter if it's Gnome or KDE or something inferior to both, as long as there's one of it and it's universal.
Once upon a time we knew what terrorism meant, and we accepted anti-terrorism laws on that basis. Now we hear of people being prosecuted for hoax 911 calls under anti-terrorism legislation because they "jeopardize the integrity of the emergency services". Does this sort of prosecution get thrown out by the judges, or will "feature creep" end up defining every socially undesirable act as terrorism? If the latter, how do we protect ourselves? As a cryptographer, I'd like to know!
Child abuse suffers the same inflation. If you don't search your 15-year-old son's bedroom for dirty magazines every day, you are "permitting him to have access to sexually explicit material", which gets you prosecuted, put on the sex offenders' register, etc. But then, the same will happen to him when he kisses his 15-year-old girlfriend...
Suppose I'm running a website that collects personal information. As a responsible citizen I have a duty to cooperate with law enforcement, especially with serious offences. [Let's leave aside for the moment the question of what governments consider as "serious"]. How can I verify which requests for access are in fact ones that I am obliged to obey? Do I have to retain a whole legal department to decide whether I have to accede to requests from Little Rock, Arkansas on the one hand or of Moscow, Russia on the other? Is there, in any jurisdiction you are aware of, some concept of "single-sourcing" of law enforcement demands so that it is easy for the person on the receiving end to determine what requests ought, and what requests ought not, to be granted?
You could try the analogy of fuses in electrical equipment. They are always unnecessary -- unless the system fails, when they save life. It is ridiculous, now, to imagine a government that will (for example) round up and intern every Semite in the country, but that is no reason not to protect against it... or at least, not to make it something that can be done at the press of a button. It happened to the Japanese, after all, 60 years ago: who knows how the technology will be used in 60 years' time? Present privacy as precautionary, prudent, like fuses. It might help.
Speaking as a software developer whose package runs under Windows (because that's where the money is) but sees that platform collapsing into an entropic mess in the not too distant future, I'm interested in ways towards liberation.
If (judging by the comments on this story) adapting your product to WINE and then doing a native port isn't a viable strategy, what is?
There are a lot of vendors like us out there and a little bit of guidance could result in a flood of Linux products.
Opera (4.0 to 6.0 inclusive) randomly forgets that I'm logged in, so I become AC halfway through doing a post or comment. Also, it needs to be told to refresh the Slashdot home page every time I run it, despite all the settings (according to the Opera people) being OK.
When I queried this with the/. gurus they said, basically, "Opera is broken so we don't support it". Given/.'s claimed desire to support things other than IE, I'd have thought that *someone* might have wanted to talk to the Opera people and tell them how to correct their program? Or even tell *us* what the incompatibility is???
Ross Anderson, in "Security Engineering" (great book) reports a better scam where the ATM was full of money and paid out whatever you asked it. You weren't likely to complain-- but more to the point, those guys had style!
One word: memory-mapped files. In a 32-bit architecture you can only memory-map small files (there is a limit of 2GB per process), so any application looking to use memory mapping to make file access faster and/or simpler (and MM is a big win on both counts) can't do it. Instead we have to write a raft of code to duplicate what the OS does better (buffering, preload, write-behind...) just in case the user wants to open a couple of medium-sized files simultaneously. With 64-bit addressing we can memory-map *all* files, regardless. It's a big win even if you only have 512MB of physical RAM.
It's too late. Office is too big and too old. Any change will wreck something else. Not least, because core aspects of its behaviour were never documented, even within MS.
Wrote my own multithreading OS for an embedded 6502. Task switches beautifully fast because only 2 or 3 registers to save. Most instructions = 1 cloxycle. Basically RISC before RISC existed.
Depends on the terms of the warranty. On the other hand, a piece of equipment that fails after N years *because of an original manufacturing defect* gives you rights against the supplier (not the manufacturer but the guys who sold you the eqpt) under UK consumer legislation, irrespective of what the warranty may say. The problem when suing the shop is normally lack of *evidence* (how do you prove that it's gone wrong today because it was mismanufactured 5 years ago?) but in this case the evidence exists.
The thing that makes you program so much better in assembler is the fact that the debugging tools aren't nearly as good.
The best code I ever wrote was when I wrote my own homebrew assembler in Basic (on a Pertec PCC2000 with ROM basic, no other OS). I didn't write a debugger, so the only way of testing a routine was to run it. You very quickly learn to write things RIGHT when any bug means rebooting and starting again!
I wrote a whole CP/M BIOS using that system, so then I had the world's only PCC2000 with CP/M. But more seriously, the code I produced there lasted 8-10 years in commercial packages I produced. With no bugs. Not one.
I'm not boasting: it was the lack of a debugger that did it.
For reasons already explained here, VLBI is a non-starter. But given good enough control, you could make a neat pair of virtual binoculars. My eyes are 10cm apart, approximately. If the left eye is fed from my own telescope and the right eye is fed from my friend's telescope 12,000 km away (diameter of the Earth), the Moon (400,000 km away) would be as three-dimensional through the "binoculars" as an object ten feet in front of my nose. With the same setup, a geosynchronous satellite would be as 3D as an object a foot away, and the International Space Station is so close that even a West Coast / East Coast telescope pair should make it nicely three-dimensional. Setting this up involves relatively little bandwidth, since you don't need real-time video speeds and you don't need phase information. You'd need to set up some clever control gear so you and your buddy can control each other's telescopes remotely. Best of all, this is an area where amateurs can really contribute because their telescopes are essentially free and their time is their own. Imagine the hellish bureaucracy in getting two institutions on the other side of the world to collaborate! And yes, I do know that 180 degrees apart wouldn't be too possible, given the light from the sun and all that; but 60 degrees should be pretty readily attainable, and that only halves the baseline so a 3D Moon should still be possible.
I don't care about 64-bit arithmetic. 64-bit pointers on the other hand... they give you enough memory that ANY hard disk can be memory-mapped. For database applications that's a big win. At present you can only memory map small files (1GB) and still need to keep the apparatus for file I/O in case people want to create something larger. But with around 8EB of virtual memory you can keep the whole DB in memory and let the OS deal with *all* the physical details. Anyone know a good Linux to start developing on, now?
Slashdot Mirror
Because one day Microsoft will ban you from using anything that resembles what you now call "Windows". In the interest of providing you a better service, of course, and nothing to do with trying to get more revenue.
Or Microsoft will die. Computer companies are short-lived things, compared to their customers.
The patch they released 21 days ago was the one that broke SHARE detection so that all our users were getting "Missing SHARE.EXE" reports from their applications. We had to patch our app to work with the patched XP.
Fortunately users are so used to Microsoft buggering things up that they accept it as a perfectly normal part of life and don't claim that we're lying to get ourselves out of trouble.
Pass legislation enabling anyone who buys a product as a result of receiving a spam to repudiate the charge on his credit card.
The card companies will start vetting their merchants more thoroughly PDQ. Money=muscle.
All that is needed is the enactment of a simple law: if you purchase any product as a result of receiving spam, you are free to repudiate the debt. In particular, if your credit card was charged, your credit card company must credit your account on receiving evidence that you made the purchase in response to a spam.
This avoids all the tedious process of law enforcement: no-one gets prosecuted, spam remains entirely legal, free speech isn't affected...
We let our users password-protect their databases. So of course they lose the passwords and we have to have a mechanism (challenge-response) to let them break in and reset the master password.
The problem is, how do we know the person asking for this service is the owner of the data? There's no way (that I can see) of both guaranteeing that a thief won't ask to have his password broken into and that a legitimate owner won't be prevented from rescuing his own data.
Some time in the last century (shortly after PCs had been invented), we created an estate agency software package called Monopoly. PC User published a review of it and illustrated it with a photograph of a certain property trading board game.
We got a letter from Waddingtons' lawyers. We wrote back to them saying that (a) we were not using "Monopoly" as the name of a game; (b) we were not implying endorsement by Waddingtons or association with their game; (c) we were not mentioning their game in any way in our documentation or marketing material.
They didn't have a leg to stand on and they didn't even reply. We carried on using the name.
If it's a registered trademark then it is registered for particular classes of goods. For example, VAX is the trademark of a vacuum cleaner, of a type of computer, and of coloured felt-tips. Three different trademark classes, three different owners, no conflict. You must not use the trademark to denote goods in those classes. In addition, you must not imply association with the trademark or promote confusion among the public. If you are using "hard radio" to mean something that is nothing to do with what HardRadio is or does, (eg. if they are using it to make people think of "hard-core radio" and you are using "hard radio" to mean radio implemented by hardware rather than software) then they don't have much of a case. If you are using it to promote something similar with or competitive to them, then they have a very good case. Remember what trademarks are for, and you won't go far wrong.
This is UK rather than US law. As far as I can see, it would be quite reasonable under UK law for someone to market a brand of too-small too-tight condoms under the name Microsoft... as long as they could show that they were using the brandname to sell the condoms and not merely to devalue the trademark of another company. Even if the case went to court, it would hinge on whether such devaluation was actually likely.
Let's go for Venus instead. Lots of raw materials to play with, and it'll be fun to design micro-organisms that have a lifecycle floating in the clouds. The bacteria from terrestrial black smokers will be really surprised to find themselves in their new location....
You're probably covered for $2-3m already, under one policy or other, if (say) you are careless crossing the road and cause a major accident.
PC risks are just as insurable - and it will be interesting to see how the insurance market prices $1m of cover for Windows systems versus Linux systems versus...
You mean, you can get the cashier at Safeway to shout out loud any word you want? Don't complain, enjoy!
Mr Ihoputa - it's a real Japanese surname.
Ms Foyame - especially if the cashier is female and cute. [there's an acute accent on the final E].
Mr Chupapoyas - especially if the cashier is male and camp.
This applies to the southern USA, but Canadians can have their own equivalent.
I want to publish my ebooks on this. Anyone any idea if it runs PalmOS or what?
There is no such thing as "the Linux GUI" and until such a thing exists no serious porting effort of commercial software takes place. I don't care which **the** Linux GUI is but I am not going to (a) run two parallel development projects or (b) bet my company on whether VHS or Beta will win in the marketplace.
It doesn't matter if it's Gnome or KDE or something inferior to both, as long as there's one of it and it's universal.
Once upon a time we knew what terrorism meant, and we accepted anti-terrorism laws on that basis. Now we hear of people being prosecuted for hoax 911 calls under anti-terrorism legislation because they "jeopardize the integrity of the emergency services". Does this sort of prosecution get thrown out by the judges, or will "feature creep" end up defining every socially undesirable act as terrorism? If the latter, how do we protect ourselves? As a cryptographer, I'd like to know!
Child abuse suffers the same inflation. If you don't search your 15-year-old son's bedroom for dirty magazines every day, you are "permitting him to have access to sexually explicit material", which gets you prosecuted, put on the sex offenders' register, etc. But then, the same will happen to him when he kisses his 15-year-old girlfriend...
Suppose I'm running a website that collects personal information. As a responsible citizen I have a duty to cooperate with law enforcement, especially with serious offences. [Let's leave aside for the moment the question of what governments consider as "serious"].
How can I verify which requests for access are in fact ones that I am obliged to obey? Do I have to retain a whole legal department to decide whether I have to accede to requests from Little Rock, Arkansas on the one hand or of Moscow, Russia on the other?
Is there, in any jurisdiction you are aware of, some concept of "single-sourcing" of law enforcement demands so that it is easy for the person on the receiving end to determine what requests ought, and what requests ought not, to be granted?
You could try the analogy of fuses in electrical equipment. They are always unnecessary -- unless the system fails, when they save life.
It is ridiculous, now, to imagine a government that will (for example) round up and intern every Semite in the country, but that is no reason not to protect against it... or at least, not to make it something that can be done at the press of a button.
It happened to the Japanese, after all, 60 years ago: who knows how the technology will be used in 60 years' time?
Present privacy as precautionary, prudent, like fuses. It might help.
Same M.O.
Speaking as a software developer whose package runs under Windows (because that's where the money is) but sees that platform collapsing into an entropic mess in the not too distant future, I'm interested in ways towards liberation.
If (judging by the comments on this story) adapting your product to WINE and then doing a native port isn't a viable strategy, what is?
There are a lot of vendors like us out there and a little bit of guidance could result in a flood of Linux products.
Opera (4.0 to 6.0 inclusive) randomly forgets that I'm logged in, so I become AC halfway through doing a post or comment. Also, it needs to be told to refresh the Slashdot home page every time I run it, despite all the settings (according to the Opera people) being OK.
/. gurus they said, basically, "Opera is broken so we don't support it". Given /.'s claimed desire to support things other than IE, I'd have thought that *someone* might have wanted to talk to the Opera people and tell them how to correct their program? Or even tell *us* what the incompatibility is???
When I queried this with the
Ross Anderson, in "Security Engineering" (great book) reports a better scam where the ATM was full of money and paid out whatever you asked it. You weren't likely to complain-- but more to the point, those guys had style!
One word: memory-mapped files. In a 32-bit architecture you can only memory-map small files (there is a limit of 2GB per process), so any application looking to use memory mapping to make file access faster and/or simpler (and MM is a big win on both counts) can't do it. Instead we have to write a raft of code to duplicate what the OS does better (buffering, preload, write-behind...) just in case the user wants to open a couple of medium-sized files simultaneously.
With 64-bit addressing we can memory-map *all* files, regardless. It's a big win even if you only have 512MB of physical RAM.
It's too late. Office is too big and too old. Any change will wreck something else. Not least, because core aspects of its behaviour were never documented, even within MS.
Wrote my own multithreading OS for an embedded 6502. Task switches beautifully fast because only 2 or 3 registers to save. Most instructions = 1 cloxycle. Basically RISC before RISC existed.
Depends on the terms of the warranty. On the other hand, a piece of equipment that fails after N years *because of an original manufacturing defect* gives you rights against the supplier (not the manufacturer but the guys who sold you the eqpt) under UK consumer legislation, irrespective of what the warranty may say. The problem when suing the shop is normally lack of *evidence* (how do you prove that it's gone wrong today because it was mismanufactured 5 years ago?) but in this case the evidence exists.
The best code I ever wrote was when I wrote my own homebrew assembler in Basic (on a Pertec PCC2000 with ROM basic, no other OS). I didn't write a debugger, so the only way of testing a routine was to run it. You very quickly learn to write things RIGHT when any bug means rebooting and starting again!
I wrote a whole CP/M BIOS using that system, so then I had the world's only PCC2000 with CP/M. But more seriously, the code I produced there lasted 8-10 years in commercial packages I produced. With no bugs. Not one.
I'm not boasting: it was the lack of a debugger that did it.
For reasons already explained here, VLBI is a non-starter. But given good enough control, you could make a neat pair of virtual binoculars. My eyes are 10cm apart, approximately. If the left eye is fed from my own telescope and the right eye is fed from my friend's telescope 12,000 km away (diameter of the Earth), the Moon (400,000 km away) would be as three-dimensional through the "binoculars" as an object ten feet in front of my nose. With the same setup, a geosynchronous satellite would be as 3D as an object a foot away, and the International Space Station is so close that even a West Coast / East Coast telescope pair should make it nicely three-dimensional.
Setting this up involves relatively little bandwidth, since you don't need real-time video speeds and you don't need phase information. You'd need to set up some clever control gear so you and your buddy can control each other's telescopes remotely.
Best of all, this is an area where amateurs can really contribute because their telescopes are essentially free and their time is their own. Imagine the hellish bureaucracy in getting two institutions on the other side of the world to collaborate!
And yes, I do know that 180 degrees apart wouldn't be too possible, given the light from the sun and all that; but 60 degrees should be pretty readily attainable, and that only halves the baseline so a 3D Moon should still be possible.
I don't care about 64-bit arithmetic. 64-bit pointers on the other hand... they give you enough memory that ANY hard disk can be memory-mapped. For database applications that's a big win. At present you can only memory map small files (1GB) and still need to keep the apparatus for file I/O in case people want to create something larger. But with around 8EB of virtual memory you can keep the whole DB in memory and let the OS deal with *all* the physical details. Anyone know a good Linux to start developing on, now?