What I read in the past was that in most States consumers that got Windows preloaded on systems couldn't sue Microsoft, because the law considered the OEM the customer, not the end consumer. That would mean that only people who purchased the upgrade or full retail version could sue Microsoft. There probably aren't enough of those to make the greedy lawyers drool enough to start a class action suit.
I bought a PC => Windows came with it => I don't want Windows => They won't let me return it => MS screwed me on ~$100.
The OEM you bought the PC from signed an agreement with MS to pay for a copy of MS Window for every PC they sell of a certain configuration. This makes the record keeping simpler for the OEM and Microsoft, and the OEM gets a better price because of it. You buy one of those PCs, and the OEM doesn't want to refund you a portion of the purchase price because it's not cheaper for them to sell you a PC without Windows. It's still the OEM that's ripping you off, not Microsoft. Microsoft is giving the OEM a discount for a valid reason. Microsoft has had problems with OEMs paying them for the software they load in the past. Their problems with IBM which came to around $12 Million, and were used as the excuse for MS not giving IBM an OEM license for Win95 untill the last moment comes to mind as an example. It's much easier for Microsoft to audit the OEM if all they have to fin out is hom many of a particular type of PC were shipped.
The OEM agrees to this because it makes thier record keeping simpler/cheaper, and they get a better price for Windows. The OEM has the option of making a different configuration available which doesn't have Windows pre loaded. However, supporting multiple configurations adds to support and manufacturing costs. For most OEM's it's probably not worth it. This is especially true with laptops where most things are integrated, and they want to be able to sell all the different configurations they do offer with Windows.
So if you want to buy a computer without windows, buy it from an OEM that will sell it that way, build your own, or just accept that you're paying probably around $30 (what the OEM likely paid for Windows) for wanting your computer set up differently than most people want theirs. I think the original Windows Refund Day was based on the Windows license agreement saying you could return Windows if you didn't agree with the license. You can always return the OS to the OEM and try and get your money back. However, I think it's the OEM, not Microsoft that owes you the money.
As for you being screwed out of $100, that figures a bit high. The OEM version, which is only to be sold with a new computer goes for $89. The local place I buy hardware at when I'm putting together a new PC or upgrading will only sell it to you if you're buying parts that might go into a new PC, such as a hard drive or motherboard. They are also perfectly willing to build you a new PC with the parts you specify, and won't charge you for the OS if you don't want it installed.
If you don't like the package one vendor is putting together, don't buy it.
Unless they changed it, you had to pay for the OS if you were using it for commercial purposes. Of course, maybe they did change it, after all they can still make their money off of the Sparc hardware. Isn't giving away IE and while charging for Windows what got MS in trouble?
Are we talking about the same MS software (windows, for example, that has gone from retail $69 to retail $189)?
The $189 retail price for Win98 is for the full version. There has NEVER been a full version of Windos available for a retail price of $69. Before making childish comments like "If so, you need to learn some math. Down is towards zero. Up is away from zero." it would be nice if you were at least objective enough to compare apples to apples.
Your OS competitiors example is another apples to oranges comparison. Red Hat and Debian aren't selling you an OS. They can't, they don't own the copyright.
I think Solaris X86 is still around $600 for commercial use. If you want to see expensive, check what a development license for some of the more popular RTOSs cost.
One of the main reasons it's hard to do Linux drivers is that Linus is against Linux having a binary driver interface. This makes it a real pain to make maintainable drivers for Linux. If you don't see drivers getting updated to work with new kernel revisions, this is probably a large part of the problem.
Re:unix is NOT an operating system
on
Is UNIX An OS?
·
· Score: 3
All of my servers run linux without X and without a monitor... am i running os less servers?
This is what's called arguing from the particular to the general. It's not logically valid. By your argument, it there is some part of the kernel that's not used on someone's system, then it's not part of the OS. Is telnet part of the OS? You don't need to have it or networking at all for that matter. Just because you don't use it on your system doesn't mean it's not part of the OS.
With Firewire every device is a busmaster. There has to be control logic in every device to know how to arbitrate to decide who can send when.
With USB, only the controller in the host can start a data transfer. When a new device is plugged into USB it tells the host controller how often the controller should poll it for new data. This kind of design makes the relatively dumb devices on USB much cheaper to implement. Think of it like having a server and a bunch of dumb terminals vs having a bunch of distributed PCs.
The above description is a bit of an oversimplification, but the savings go beyond hardware costs. Simpler hardware is easier to debug. Since the USB controller is being imbedded in chipsets, you have fewer implementations which you need to test against in order to verify your product will work when it gets to the customer.
Sounds like Firewire suits your needs just fine. It seems much more likely that USB 2.0 will be used in lower cost products where saving a few bucks on the controller logic is important. When scanners and printers are selling for under $100, every buck the hardware developers can save is important.
The features that make Firewire superior to USB as the same data rate, such as peer to peer communications, also make the logic in the devices much more expensive. It's just not cost effective to make USB mice and keyboards.
I keep hearing the question of why do we need USB 2.0 when we have Firewire. The average consumer doesn't ask why they can't use Firewire. They see some new gadget they want to hook to their PC (maybe a fast CD-RW drive) and ask why they can't plug it into one of the thingys in the back of their computer. (I'm know that there are currently some CD-RW drives that work with USB 1, but when you get to higher X speeds 12Mbps isn't enough.) I don't know how many times I've had to explain to new computer users what all the connectors on the back of their PC are for. Let me explain with an example.
Your mother comes to you and says that she wants one of those CD-RW thingys her friend in her card club has. Her friend told her to get at least a 12X/8X/32X version. She knows she needs it to be that fast, and wrote down the numbers be sure she got it right. She has a PC she has a PC she bought at Best Buy (the same friend recomended it to her) which has absolutely no drive bays left, and you can't remove her CD drive because she wants to make direct CD to CD copies. After all copying the CD to the hard drive and back out to the CD-RW would require finding the CD image back on the hard drive after it was written there, and things are always "disappearing" on her hard drive.
So, she starts asking you questions, because although she believes everything here card club buddie says is true, everything you say is suspect. She still remembers you lied to her when you were four. First she asks why she can't connect the funny thing with nothing plugged into it (9 pin serial port). You patiently tell her that the serial port can't be used for that. She asks what's it used for then, and you reply that they are used for mice and modems. This of course confirms her belief that you don't really know what you are talking about. After all, the mouse is already plugged into the little port which has a mouse right next to it (PS/2 port) and for the modem, you just plug the phone line into the phone jack on the back (internal modem). Even she knows these things. She also knows that her printer hooks up to that bigger port on the back, but that the scanner gets plugged in first, and the printer gets plugged into the back of the scanner plug (parallel port scanner). Now try telling her that she really needs this Firewire thing to plug her new CD-RW into.
Wouldn't it be much easier to have just one kind of port which could handle just about anything your average user would want to attach. USB 1 and Firewire together could meet those needs, but why have two types of ports, requiring two controllers in the system. USB 2.0 delivers the needed bandwidth, while it's relative simplicity of design allows devices to be lower cost.
It is true that USB can't offer everything that Firewire can. It's not designed for peer to peer transfers. For you people who are more familiar with SCSI terminology, devices cannot be initiators, only targets. There also some other differences, but for the average consumer, who just wants to hook up some neet new peripheral, it really doesn't make a difference.
USB 2.0 insn't some Intel conspiracy to only use standards they came up with, or maybe it is, I don't really care. It fills a need that exhists in a cost efective manner. Sounds like good engineering to me. Maybe Firewire will find it's niche. I heard that Western Digital came out with a Firewire drive, and Firewire's still a lot cheaper than FibreChannel. Or maybe it will end up like the Apple Desktop Bus, something that really seemed like a good idea at the time, but just never took off. Apple sure seems to have a lot of those.
Firewire is not only faster than USB, it's more versatile, and more expensive. Firewire mice and keyboards just aren't very practical. Unfortunately, the 12Mb/sec bandwidth of USB 1.0 is a little slow for some peripherals such as scanners and CD-RW drives. Putting Firewire on PCs to meet this need means more ports of different types to confuse the average user. It also means additional expense of having both a USB controller and Firewire controller in the same system. With USB 2.0, the additional cost is relatively low, and it keeps things simpler for the user. Maybe Firewire will continue to increase it's bandwidth and still take off for higher end devices. So far, the only ones I see pushing it are Apple and Sony. If it's such a good general purpose bus, where's the iLink (Firewire) ports on the Playstation 2. Of course, I've never seen a Playstation 2, so for all I know it could have Firewire.
Plus, when you've got a few billion in cash, it's not a bad idea to have a few products in your back pocket waiting for hte right time to release.
It's hard to think of a better reason than this one. There's nothing marketing people seem to hate more than making a decision. If you can afford to develop products you may want to have in the future, it sure reduces the time to market when you finally decide what you want.
but the Windows operating system has plenty of architectural flaws
I should probably leave this alone, but what do you consider to be Windows' architectural flaws. I can readily agree that Win9X has it's share, but I courious what archetectural issues you have with the NT/2000 kernel? I'm not saying NT is bug free, but I have trouble calling Linux's kernel architecturally better at this point. I don't have a lot of experience with the Linux kernel, but a lot of things seem to be hacked together as it has evolved. Some things I would expect in a well designed kernel are good kernel threads support, the ability to map user memory into kernel space, and a device driver interface. A consistent device driver interface is especially important when the OS is evolving as quickly as Linux is right now. At least it's important to those of us writing the device drivers.
I understand people not liking Windows because of stability problems.
I can understand people not liking it because it's closed source.
I can understand people not liking it because it's sold by Microsoft.
But the architecture of NT is pretty dan impressive.
The price of hardware tends to drop quickly, however a lot of the costs of implementing this will be in software development. It's not like the developers can make their money back by selling lots of licenses. The work will a specific application for a specific customer. Software engineering time doesn't appear to be getting much cheaper, and I doubt it'll become a open source project.
The French government is not simply objecting to these items being sold to French citizens. They are objecting to their citizens being able to see the items displayed on the site at all. I don't thing the French government has any right to tell Yahoo that they need to find a way to block French users from seeing something on a server located in the United States. The server is outside their jurisdiction. The only reason Yahoo has to listen to the French government about this is becase they have a division in France. Their French site www.yahoo.fr complies with French law. If the French government doesn't want their citizens accessing those sites, then they should force French ISPs to not allow their citizens access to them.
I'm pretty sure Yahoo has assets in France such as servers, money in French bank accounts. If Yahoo ignores the government, their employees in France could possibly go to jail, and their assets there could be seized. This would be very bad for Yahoo, but it wouldn't be very good for France economically. It would likely discourage net based companies from having a presence in France which would have an effect on their tax base.
The servers are in the United States. The business isn't taking place in france, it's taking place in the US. If the French government doesn't want it's citizens to access those servers, then the French government should find some domestic way to keep their citizens from doing so.
Perhaps a better solution would be for all French users to register with the goverment, so that Yahoo could block their IPs.
A better solution would be for the French government to find a way to prevent French citizens from reaching those sites domestically. They should not be trying to force a company in a different country to enforce their own national laws. Yahoo has abided by their laws on their site which was created for the French. The servers for the site in question is not in France. If the French government wants their citizens banned from accessing content on them, then it's the French government's problem, not Yahoo's. If the French can't block the user's domestically, then they should be able to approach Yahoo and see if Yahoo would be willing to implement some method of blocking their users, but the government better be willing to pay for it.
The French government seems to have a problem understanding other countries don't have to enforce their domestic laws for them. Our government here in the United States also suffers from this delusion from time to time.
If the French government can't deal with this then Yahoo should probably sell their French assets. If they don't have any assets in France, there's probably not much the French government can do if Yahoo choses not to comply with their wishes. If the French govenment doesn't want to play nice, then let them play by themselves.
I wholeheartedly agree that the police sholdn't be able to search someone without probable cause. However, his actions pretty much amounted to probable cause in this case. They probably also amounted to disturbing the peace for which he could be arrested, and therefore searched. You have a constitiuional right against unreasonable searches. There's no constitutional right to be an asshole.
For the most part I think the market has decided that 32-bit Windows drivers are much better than 16bit DOS ones. However you will always run into the case where some hardware vendor already has a bunch of old 16-bit DOS code which will work for their new hardware. The engineers probably intend to write a 32-bit driver for it eventually, but the current code will work long enough to test the hardware, and as soon as something kinda works, marketing starts shipping it to customers. Try convincing marketing you need to completely rewrite your driver to make it better before you can ship a product. For a lot of Engineers the hardware is the interesting part, they don't really want to work on the driver either. They want to play with the next generation of the hardware.
The result of this is that consumers buy new hardware, which has drivers that work under DOS, and if it doesn't work, they blame Microsoft. After all, if they call tech support they will say it worked under Windows98, just like they said it would. It was that evil Microsoft that went and changed things.
Then again, maybe the hardware vendors have found the drivers are more stable in the 16bit DOS world as opposed to the all encompassing Windows API's where applications and browsers are part of the OS?
I can see how it could be easier for developers to hack together drivers under DOS, however stability is definately not one of the benefits. With DOS drivers you are much more likely to have device conflicts and cause the system to be less stable. Of course a developer can still write a crappy NT driver which will conflict with other hardware as well as crash the system. It's just a lot more obvious that you're doing things in the driver that you shouldn't be doing. You also have a lot more flexability writing drivers with multiple threads and asynchronous I/O which DOS can't provide.
The real reason for hiding the real mode DOS is marketing. Microsoft can now say even more convincingly: "Look mom, no DOS!".
Actually, I think the purpose of marketing Win9X as not haveing DOS is to get users and developers to quit using it so Microsoft can migrate users to Win2k. It's really in Microsoft's best interest to continue developing both the NT and Win9X kernels, however they need to get the users to buy into it, which has taken a few more revisions of Win9X than were planned.
You act like this is some surprise Microsoft came up with to impede competition.
DOS is not just a lower operating system--it's a basic environment that can be entirely overwritten by whatever code happens to run underneath it.
This may be a surprise to you, but if you want to be able to overwrite OS functionality, maybe a consumer OS designed for the masses isn't the right OS for you. Microsoft has been telling users for years that this is the direction they are headed. They need to do this in order to migrate their users to Win2k, which will offer users a lot in terms of reliability. Not only does Microsoft get their users a more stable OS, but they would no longer have to support two development paths. That would be a huge savings for them.
The number of users that benefit from having DOS is very small. Not only that, but poorly written software that can directly access the hardware can do a lot of damage, and there's a lot of poorly written software out there. Old DOS apps, direct hardware access represent support nighmares for Microsoft. They likely cost them more to support than they make selling the OS to the customer. The only reason this aspect of backward compatability has lasted this long is to avoid pissing off too many customers. They've tried many times to get their customer base away from DOS, and each time they've reduced the relaince on it a little more.
I really doubt that the purpose of this is to specificly harm competition. I doubt Microsoft is going to be real upset if it hurts any competing products, but Microsoft does have a valid reason for doing this. It's really not Microsoft's responsibility to make sure they don't do anything to accidently harm their competitors.
They just hid it better, of course. They already have an OS without DOS. WinNT/Win2000 is the direction Microsoft has been dragging their users for a long time. They reall just want people (especially developers) to quit using DOS mode, so they can quit supporting two OS that are very similar on the surface, but considerable different in the kernel. They would love to dump Win9X and significantly cut down on their development costs, as well as deliver a more stable product to their customers. However, their installed customer base isn't willing to give up their backward compatability. So Microsoft hides DOS a little better every version, and tries to only break a few "important" legacy apps each time. Eventually, they will drag their customers away from the crap that is DOS to NT/Win2000. If Microsoft really had total control in the market they would have forced their customers to NT years ago. Just like any other vendor with a large installed customer base, they are a slave to backward compatability.
The problem is that many game developers are still writing games that won't load or run properly on Win2k. One of the things Microsoft is trying to do with ME is force developers to start writing software that's compatible with Win2k as well as the Win9X series.
Choice is a good thing, yes. The problem with Unix isn't choice.
Choice is a good thing only to a certain extent. On a desktop OS, if everyone chooses to configure how their desktop operates, it's pure hell for computer support. If your're trying to write an application for everyone in your company to use, and everyone wants to use a different Linux distribution, you're likely to run into problems. The problems may only be making sure everyone has the same libraries and kernel revision, but it's still a pain. It's nice to have choices, but unless you don't have to interact with other people, the choises have to be considerably limited. If you want to set your home Linux system up differently than everyone else, great. The problem occurrs when someone else needs to use that system. That's when a standard way of doing things has to be created and enforced.
Slashdot Mirror
What I read in the past was that in most States consumers that got Windows preloaded on systems couldn't sue Microsoft, because the law considered the OEM the customer, not the end consumer. That would mean that only people who purchased the upgrade or full retail version could sue Microsoft. There probably aren't enough of those to make the greedy lawyers drool enough to start a class action suit.
I bought a PC => Windows came with it => I don't want Windows => They won't let me return it => MS screwed me on ~$100.
The OEM you bought the PC from signed an agreement with MS to pay for a copy of MS Window for every PC they sell of a certain configuration. This makes the record keeping simpler for the OEM and Microsoft, and the OEM gets a better price because of it. You buy one of those PCs, and the OEM doesn't want to refund you a portion of the purchase price because it's not cheaper for them to sell you a PC without Windows. It's still the OEM that's ripping you off, not Microsoft. Microsoft is giving the OEM a discount for a valid reason. Microsoft has had problems with OEMs paying them for the software they load in the past. Their problems with IBM which came to around $12 Million, and were used as the excuse for MS not giving IBM an OEM license for Win95 untill the last moment comes to mind as an example. It's much easier for Microsoft to audit the OEM if all they have to fin out is hom many of a particular type of PC were shipped.
The OEM agrees to this because it makes thier record keeping simpler/cheaper, and they get a better price for Windows. The OEM has the option of making a different configuration available which doesn't have Windows pre loaded. However, supporting multiple configurations adds to support and manufacturing costs. For most OEM's it's probably not worth it. This is especially true with laptops where most things are integrated, and they want to be able to sell all the different configurations they do offer with Windows.
So if you want to buy a computer without windows, buy it from an OEM that will sell it that way, build your own, or just accept that you're paying probably around $30 (what the OEM likely paid for Windows) for wanting your computer set up differently than most people want theirs. I think the original Windows Refund Day was based on the Windows license agreement saying you could return Windows if you didn't agree with the license. You can always return the OS to the OEM and try and get your money back. However, I think it's the OEM, not Microsoft that owes you the money.
As for you being screwed out of $100, that figures a bit high. The OEM version, which is only to be sold with a new computer goes for $89. The local place I buy hardware at when I'm putting together a new PC or upgrading will only sell it to you if you're buying parts that might go into a new PC, such as a hard drive or motherboard. They are also perfectly willing to build you a new PC with the parts you specify, and won't charge you for the OS if you don't want it installed.
If you don't like the package one vendor is putting together, don't buy it.
Unless they changed it, you had to pay for the OS if you were using it for commercial purposes. Of course, maybe they did change it, after all they can still make their money off of the Sparc hardware. Isn't giving away IE and while charging for Windows what got MS in trouble?
Are we talking about the same MS software (windows, for example, that has gone from retail $69 to retail $189)?
The $189 retail price for Win98 is for the full version. There has NEVER been a full version of Windos available for a retail price of $69. Before making childish comments like "If so, you need to learn some math. Down is towards zero. Up is away from zero." it would be nice if you were at least objective enough to compare apples to apples.
Your OS competitiors example is another apples to oranges comparison. Red Hat and Debian aren't selling you an OS. They can't, they don't own the copyright.
I think Solaris X86 is still around $600 for commercial use. If you want to see expensive, check what a development license for some of the more popular RTOSs cost.
Although this would help drive the Packard Bells of the industry out faster, it's biggest effect will be to raise the price of computers.
As for those of us who upgrade our hardware, device drivers, OS, or even install service packs, our warranty would be void.
I'm from Pennsylvania, and I thought it was funny.
One of the main reasons it's hard to do Linux drivers is that Linus is against Linux having a binary driver interface. This makes it a real pain to make maintainable drivers for Linux. If you don't see drivers getting updated to work with new kernel revisions, this is probably a large part of the problem.
All of my servers run linux without X and without a monitor... am i running os less servers?
This is what's called arguing from the particular to the general. It's not logically valid. By your argument, it there is some part of the kernel that's not used on someone's system, then it's not part of the OS. Is telnet part of the OS? You don't need to have it or networking at all for that matter. Just because you don't use it on your system doesn't mean it's not part of the OS.
With Firewire every device is a busmaster. There has to be control logic in every device to know how to arbitrate to decide who can send when.
With USB, only the controller in the host can start a data transfer. When a new device is plugged into USB it tells the host controller how often the controller should poll it for new data. This kind of design makes the relatively dumb devices on USB much cheaper to implement. Think of it like having a server and a bunch of dumb terminals vs having a bunch of distributed PCs.
The above description is a bit of an oversimplification, but the savings go beyond hardware costs. Simpler hardware is easier to debug. Since the USB controller is being imbedded in chipsets, you have fewer implementations which you need to test against in order to verify your product will work when it gets to the customer.
Sounds like Firewire suits your needs just fine. It seems much more likely that USB 2.0 will be used in lower cost products where saving a few bucks on the controller logic is important. When scanners and printers are selling for under $100, every buck the hardware developers can save is important.
The features that make Firewire superior to USB as the same data rate, such as peer to peer communications, also make the logic in the devices much more expensive. It's just not cost effective to make USB mice and keyboards.
I keep hearing the question of why do we need USB 2.0 when we have Firewire. The average consumer doesn't ask why they can't use Firewire. They see some new gadget they want to hook to their PC (maybe a fast CD-RW drive) and ask why they can't plug it into one of the thingys in the back of their computer. (I'm know that there are currently some CD-RW drives that work with USB 1, but when you get to higher X speeds 12Mbps isn't enough.) I don't know how many times I've had to explain to new computer users what all the connectors on the back of their PC are for. Let me explain with an example.
Your mother comes to you and says that she wants one of those CD-RW thingys her friend in her card club has. Her friend told her to get at least a 12X/8X/32X version. She knows she needs it to be that fast, and wrote down the numbers be sure she got it right. She has a PC she has a PC she bought at Best Buy (the same friend recomended it to her) which has absolutely no drive bays left, and you can't remove her CD drive because she wants to make direct CD to CD copies. After all copying the CD to the hard drive and back out to the CD-RW would require finding the CD image back on the hard drive after it was written there, and things are always "disappearing" on her hard drive.
So, she starts asking you questions, because although she believes everything here card club buddie says is true, everything you say is suspect. She still remembers you lied to her when you were four. First she asks why she can't connect the funny thing with nothing plugged into it (9 pin serial port). You patiently tell her that the serial port can't be used for that. She asks what's it used for then, and you reply that they are used for mice and modems. This of course confirms her belief that you don't really know what you are talking about. After all, the mouse is already plugged into the little port which has a mouse right next to it (PS/2 port) and for the modem, you just plug the phone line into the phone jack on the back (internal modem). Even she knows these things. She also knows that her printer hooks up to that bigger port on the back, but that the scanner gets plugged in first, and the printer gets plugged into the back of the scanner plug (parallel port scanner). Now try telling her that she really needs this Firewire thing to plug her new CD-RW into.
Wouldn't it be much easier to have just one kind of port which could handle just about anything your average user would want to attach. USB 1 and Firewire together could meet those needs, but why have two types of ports, requiring two controllers in the system. USB 2.0 delivers the needed bandwidth, while it's relative simplicity of design allows devices to be lower cost.
It is true that USB can't offer everything that Firewire can. It's not designed for peer to peer transfers. For you people who are more familiar with SCSI terminology, devices cannot be initiators, only targets. There also some other differences, but for the average consumer, who just wants to hook up some neet new peripheral, it really doesn't make a difference.
USB 2.0 insn't some Intel conspiracy to only use standards they came up with, or maybe it is, I don't really care. It fills a need that exhists in a cost efective manner. Sounds like good engineering to me. Maybe Firewire will find it's niche. I heard that Western Digital came out with a Firewire drive, and Firewire's still a lot cheaper than FibreChannel. Or maybe it will end up like the Apple Desktop Bus, something that really seemed like a good idea at the time, but just never took off. Apple sure seems to have a lot of those.
Firewire is not only faster than USB, it's more versatile, and more expensive. Firewire mice and keyboards just aren't very practical. Unfortunately, the 12Mb/sec bandwidth of USB 1.0 is a little slow for some peripherals such as scanners and CD-RW drives. Putting Firewire on PCs to meet this need means more ports of different types to confuse the average user. It also means additional expense of having both a USB controller and Firewire controller in the same system. With USB 2.0, the additional cost is relatively low, and it keeps things simpler for the user. Maybe Firewire will continue to increase it's bandwidth and still take off for higher end devices. So far, the only ones I see pushing it are Apple and Sony. If it's such a good general purpose bus, where's the iLink (Firewire) ports on the Playstation 2. Of course, I've never seen a Playstation 2, so for all I know it could have Firewire.
Plus, when you've got a few billion in cash, it's not a bad idea to have a few products in your back pocket waiting for hte right time to release.
It's hard to think of a better reason than this one. There's nothing marketing people seem to hate more than making a decision. If you can afford to develop products you may want to have in the future, it sure reduces the time to market when you finally decide what you want.
but the Windows operating system has plenty of architectural flaws
I should probably leave this alone, but what do you consider to be Windows' architectural flaws. I can readily agree that Win9X has it's share, but I courious what archetectural issues you have with the NT/2000 kernel? I'm not saying NT is bug free, but I have trouble calling Linux's kernel architecturally better at this point. I don't have a lot of experience with the Linux kernel, but a lot of things seem to be hacked together as it has evolved. Some things I would expect in a well designed kernel are good kernel threads support, the ability to map user memory into kernel space, and a device driver interface. A consistent device driver interface is especially important when the OS is evolving as quickly as Linux is right now. At least it's important to those of us writing the device drivers.
I understand people not liking Windows because of stability problems.
I can understand people not liking it because it's closed source.
I can understand people not liking it because it's sold by Microsoft.
But the architecture of NT is pretty dan impressive.
The price of hardware tends to drop quickly, however a lot of the costs of implementing this will be in software development. It's not like the developers can make their money back by selling lots of licenses. The work will a specific application for a specific customer. Software engineering time doesn't appear to be getting much cheaper, and I doubt it'll become a open source project.
It's just not that simple.
The French government is not simply objecting to these items being sold to French citizens. They are objecting to their citizens being able to see the items displayed on the site at all. I don't thing the French government has any right to tell Yahoo that they need to find a way to block French users from seeing something on a server located in the United States. The server is outside their jurisdiction. The only reason Yahoo has to listen to the French government about this is becase they have a division in France. Their French site www.yahoo.fr complies with French law. If the French government doesn't want their citizens accessing those sites, then they should force French ISPs to not allow their citizens access to them.
I'm pretty sure Yahoo has assets in France such as servers, money in French bank accounts. If Yahoo ignores the government, their employees in France could possibly go to jail, and their assets there could be seized. This would be very bad for Yahoo, but it wouldn't be very good for France economically. It would likely discourage net based companies from having a presence in France which would have an effect on their tax base.
The servers are in the United States. The business isn't taking place in france, it's taking place in the US. If the French government doesn't want it's citizens to access those servers, then the French government should find some domestic way to keep their citizens from doing so.
Perhaps a better solution would be for all French users to register with the goverment, so that Yahoo could block their IPs.
A better solution would be for the French government to find a way to prevent French citizens from reaching those sites domestically. They should not be trying to force a company in a different country to enforce their own national laws. Yahoo has abided by their laws on their site which was created for the French. The servers for the site in question is not in France. If the French government wants their citizens banned from accessing content on them, then it's the French government's problem, not Yahoo's. If the French can't block the user's domestically, then they should be able to approach Yahoo and see if Yahoo would be willing to implement some method of blocking their users, but the government better be willing to pay for it.
The French government seems to have a problem understanding other countries don't have to enforce their domestic laws for them. Our government here in the United States also suffers from this delusion from time to time.
If the French government can't deal with this then Yahoo should probably sell their French assets. If they don't have any assets in France, there's probably not much the French government can do if Yahoo choses not to comply with their wishes. If the French govenment doesn't want to play nice, then let them play by themselves.
I wholeheartedly agree that the police sholdn't be able to search someone without probable cause. However, his actions pretty much amounted to probable cause in this case. They probably also amounted to disturbing the peace for which he could be arrested, and therefore searched. You have a constitiuional right against unreasonable searches. There's no constitutional right to be an asshole.
For the most part I think the market has decided that 32-bit Windows drivers are much better than 16bit DOS ones. However you will always run into the case where some hardware vendor already has a bunch of old 16-bit DOS code which will work for their new hardware. The engineers probably intend to write a 32-bit driver for it eventually, but the current code will work long enough to test the hardware, and as soon as something kinda works, marketing starts shipping it to customers. Try convincing marketing you need to completely rewrite your driver to make it better before you can ship a product. For a lot of Engineers the hardware is the interesting part, they don't really want to work on the driver either. They want to play with the next generation of the hardware.
The result of this is that consumers buy new hardware, which has drivers that work under DOS, and if it doesn't work, they blame Microsoft. After all, if they call tech support they will say it worked under Windows98, just like they said it would. It was that evil Microsoft that went and changed things.
Then again, maybe the hardware vendors have found the drivers are more stable in the 16bit DOS world as opposed to the all encompassing Windows API's where applications and browsers are part of the OS?
I can see how it could be easier for developers to hack together drivers under DOS, however stability is definately not one of the benefits. With DOS drivers you are much more likely to have device conflicts and cause the system to be less stable. Of course a developer can still write a crappy NT driver which will conflict with other hardware as well as crash the system. It's just a lot more obvious that you're doing things in the driver that you shouldn't be doing. You also have a lot more flexability writing drivers with multiple threads and asynchronous I/O which DOS can't provide.
The real reason for hiding the real mode DOS is marketing. Microsoft can now say even more convincingly: "Look mom, no DOS!".
Actually, I think the purpose of marketing Win9X as not haveing DOS is to get users and developers to quit using it so Microsoft can migrate users to Win2k. It's really in Microsoft's best interest to continue developing both the NT and Win9X kernels, however they need to get the users to buy into it, which has taken a few more revisions of Win9X than were planned.
You act like this is some surprise Microsoft came up with to impede competition.
DOS is not just a lower operating system--it's a basic environment that can be entirely overwritten by whatever code happens to run underneath it.
This may be a surprise to you, but if you want to be able to overwrite OS functionality, maybe a consumer OS designed for the masses isn't the right OS for you. Microsoft has been telling users for years that this is the direction they are headed. They need to do this in order to migrate their users to Win2k, which will offer users a lot in terms of reliability. Not only does Microsoft get their users a more stable OS, but they would no longer have to support two development paths. That would be a huge savings for them.
The number of users that benefit from having DOS is very small. Not only that, but poorly written software that can directly access the hardware can do a lot of damage, and there's a lot of poorly written software out there. Old DOS apps, direct hardware access represent support nighmares for Microsoft. They likely cost them more to support than they make selling the OS to the customer. The only reason this aspect of backward compatability has lasted this long is to avoid pissing off too many customers. They've tried many times to get their customer base away from DOS, and each time they've reduced the relaince on it a little more.
I really doubt that the purpose of this is to specificly harm competition. I doubt Microsoft is going to be real upset if it hurts any competing products, but Microsoft does have a valid reason for doing this. It's really not Microsoft's responsibility to make sure they don't do anything to accidently harm their competitors.
They just hid it better, of course. They already have an OS without DOS. WinNT/Win2000 is the direction Microsoft has been dragging their users for a long time. They reall just want people (especially developers) to quit using DOS mode, so they can quit supporting two OS that are very similar on the surface, but considerable different in the kernel. They would love to dump Win9X and significantly cut down on their development costs, as well as deliver a more stable product to their customers. However, their installed customer base isn't willing to give up their backward compatability. So Microsoft hides DOS a little better every version, and tries to only break a few "important" legacy apps each time. Eventually, they will drag their customers away from the crap that is DOS to NT/Win2000. If Microsoft really had total control in the market they would have forced their customers to NT years ago. Just like any other vendor with a large installed customer base, they are a slave to backward compatability.
The problem is that many game developers are still writing games that won't load or run properly on Win2k. One of the things Microsoft is trying to do with ME is force developers to start writing software that's compatible with Win2k as well as the Win9X series.
Choice is a good thing, yes. The problem with Unix isn't choice.
Choice is a good thing only to a certain extent. On a desktop OS, if everyone chooses to configure how their desktop operates, it's pure hell for computer support. If your're trying to write an application for everyone in your company to use, and everyone wants to use a different Linux distribution, you're likely to run into problems. The problems may only be making sure everyone has the same libraries and kernel revision, but it's still a pain.
It's nice to have choices, but unless you don't have to interact with other people, the choises have to be considerably limited. If you want to set your home Linux system up differently than everyone else, great. The problem occurrs when someone else needs to use that system. That's when a standard way of doing things has to be created and enforced.