I'll tell you -- the MS/Intel PC 99 spec specifically overrides the previous PCI spec. It's now up to the OS to allocate PCI resources, not the BIOS, although companies can chose to have a legacy mode setting.
This, and WinModems, are just the tip of the Iceburg. Microsoft is trying to turn the PC clone into the Windows computer. They've disposed of serveral critical pieces of the IBM PC AT Clone - text mode, ISA slots, and the joystick port.
It's starting to look like a PC99 will get a Windows logo, but won't even boot MS-DOS or any other OS. RMS is right on here - if Microsoft were to get away with this, it would be far worse than things like "secret" DOC file formats. --
Just as a minor correction, Apple does compete with other Mac developers. Try to buy an iMac without paying the ClarisWorks "tax". They're used to be four or five other "works" packages for the Mac - Apple drove them under.
There's numerous other examples - Final Cut, WebObjects. I think the key is that there's no comprehensive competition like there is with Microsoft. I seriously doubt there is one Windows vendor whose #1 competition is not Microsoft themselves (well, Adobe).
1) Netscape doesn't yet have a componentized browser. Other companies such as Lotus, Quicken, and AOL are using the IE component, so the specs may be open enough to reverse engineer (then again, they may not.)
2) There's tons of dohickeys that plug into the Explorer shell. There's also been rip-and-replace alternative shells in the past. What's LiteStep trying to do.
3) Every Windows application manipulates the file associateions, Microsoft or not. It's a operating system design flaw (and one that KDE also seems to have). Look at Macintosh file types for a better implementation.
The big problem with Microsoft's interfaces is not that their secret (although it's conventional wisdom, is it true?), but that by that Microsoft only implements a new API along with a new product. They therefore get the time to market advantage.
As a contrary approach, look at Apple OpenDoc. They released a API, but then refused to use internally at Claris or in the Finder. As a third party vendor, the concern that an API is going to stick around for a long time is certainly valid, and when MS develops something like OLE, at least they use it themselves. --
According to my understanding, a user can monitor what APIs are being called by an application. Someone did catch Microsoft Office using secret APIs, but that was in 1991 or something, I have not heard about a similar scandal since.
Contrary to this assertation, one reason Microsoft has gotten into corporations favor is that every product they create has a published API. Real Media has a player, Microsoft has a Media Player SDK. You probably could write a complete application suite on top of the published API for MS Office, and so on.
The only API which I understand is "secret" is the low-level WinNT kernal API. But do any MS user space applications call this API? None that I've heard of.
Many, many software companies write fully functional non-buggy Windows applications. Are they hiring ex-Microsoft employees who snuck documentation out?
Not that I'm not saying MS is not a legal monopoly and shouldn't be penalized. Just the idea that their APIs are some bastion of secrecy (like IBM's were in the 60s) doesn't necessarily hold water. --
Every attempt at a combination Game console/Computer has not been that successful. Think Coleco Adam, Atari XE, Amiga CD, Sega Saturn plus Internet add on, Pippen, etc.
That's not to say it's a bad idea. It's just hard to design a game console (which has a shelf life of 2-5 years) that can keep up with PCs that double in speed every 18 months. --
What's wrong with WinCE in a game console? (Other than the words "Microsoft" and "Windows" that is.)
Seems to me you could have a number of interesting convergance applications, beyond web browsing. Kind of like the Atari 400 or Commodore 64 -- primarily game machines, but you could also balance your budget or whatever.
Note that in most environments, the "power user" who hacking his registry or trying to install Linux on his work machine is the worst kind of user, and the kind user that generates the most support headaches.
A real power user is someone who's writing Access (or Unix equivlant) databases or HTML/Javascript and so on. It's not someone whose trying to change his system configuration or test out some cool downloads he's found. Those people are trying to get the company to subsidize (through hardware and support) their computer hobby.
Most corporations either do disallow this behavior or would if they could. Linux/Unix is the ultimate solution in this regard simply because root is restricted enough to give these folks little or no opportunity to hack on their own configuation.
Note that I'm not expressing contempt for the average user trying to do their job here - only the special case. --
You make an extremely valid point regarding the difference between a home ("mom") system and a system that functions well in a corporate network environment. (One of the huge problems with Windows 9x is that it's trying to satisfy both these markets at the same time.)
However, in this context, Linux currently comes out behind in some respects. Linux (and Unix in general) is still very based on the host-terminal model, and therefor provides excellent remote administration capabilties, but limited capabilities to admin a group of hosts.
But in a client situation, corporations don't necessarily want better administrative tools, what's needed is less administration altogether.
Currently there's a lack of system management tools avalible, and if you did have 1000 odd Linux desktop machines, you couldn't even use something as rudimentary as Windows system policies to use default settings. In fact, as far as I can tell, you are limited to using a local account database on each machine, unless you buy Caldera's NDS module. (I could be missing NIS here.) When the #1 user complaint has traditionally been password management (too many, not syncronized, etc), this is a big problem.
So, while it may seem like a panecea to be able to telnet into Joe User's box and install or fix his word processor, what's really needed is a network-oriented system where such actions are unecessary on a user-by-user basis. X-Terminals and Java NCs have pretty much fallen by the wayside -- Maybe that's where the network admins should look, rather than trying to give the "dumb user" population Unix.
Don't forget that a good portion of the "Windows Tax" is actually end user technical support. The actual cost of Windows-with-no-support to OEM customers is only $25.
Any preloaded Linux setup is going to have a "RedHat tax" or a "Corel tax", just to pay for the tech support. Sure a clone shop could give you Linux and send you to the newsgroups when you have a problem, but don't expect Dell or Compaq to do that.
Admittedly, that $25 is significant with a $400 computer, but fogeys like myself still think computers cost $2000, and that $25 is probably insignificant in that scope.
You've hit the nail on the head right there. Various projects are making very good progress on developing a fully functional user environment for Linux, but it still seems that the system administration tools are lacking, when in fact the system admin (WinGeek) audience is probably more likely download or buy Linux than a normal user.
Someone above pointed out that there are Linux users that don't want to be Unix users, and that certainly is a valid market out there. The problem is right now, you can't be sure that the system admin tools that exist (RedHat stuff, LinuxConf) are actually working properly without breaking the "legacy" (from their point of view) unix stuff.
On a Windows or Mac system, after you've made a configuration change, 99.5% of the time you don't need to go into the configuration database to make sure that it's right. And if you do, it's rightfully labeled as a bug. Linux needs to strive for this level of functionality.
The only tools which sounds like their close are YaST (haven't tried it yet) and IBM AIX's thing.
(MasOSX may have a good solution -- A binary config database that you can dump out to standard Unix config files, which you then edit and import back.)
NTSC broadcasted over the airways will be obsolete, but your TV set will not be. You'll be able to buy and inexpensive HDTV tuner box and hook it up, much like people do with cable boxes now.
Note that most "High Definition" TV broadcast will actually be at the same resolution as normal NTSC, there'll just be more (free) channels over the airwaves. --
Well, if you look at the average MS Word user, they probably only use 20% of the feature set.
But, if you look across a medium or large organization, you'd probably find that 95% of Word's features are actively being used.
Sure a cheaper, more lightweight, more targeted solution would solve most peoples needs most of the time, but for the more advanced features, you'd have to go with other tools. This creates an integration and deployment nightmare when you are trying to support an organizational "standard".
In short, IT departments that pay through the nose for MS Office know what they're getting. There's been cheaper/lightweight solutions many times before (Volkswriter!), but they've never flown. --
I recall talking to an Netscape sales rep soon after Navigator 1.1 came out. She said that Netscape wanted to give the software away for free, because it was free advertising for the server products. But big accounts kept coming and saying that they wouldn't deploy it unless they could get a support contract. So Netscape "had to" start site licescing the thing. Plausable enough to be beliveable.
There were still many many dolts that bought the shrinkwrapped Netscape box rather than download it.
Yeah, if you think it's bad now, just wait a couple months for the big 2000 PR bulldozer. Should knock Linux right off ZDNet.
Microsoft knows it's in a bit of jam with the Win2000 server upgrade. It contains an enormous number of network services that are untested and poorly understood. IT shops (who contrary to/. legend are generally happy with NT4) are going to be understandably wary about Win2000.
In fact, many NT shops are still tying their networks together with NBT broadcasts and have a lot of work and brainpower to accumulate to implement Active Directory.
In short, there's a good chance that most of the installed base won't upgrade, much like most of the Novell 3.x base hiccuped when 4.x came out. This opens the door wide open for Linux, if it can act enough like an NT clone from the NOS perspective.
Well maybe Gates knows something not many other people do -- Microsoft Office has gotten pretty much every feature that could be possibly be devised. Pretty soon, someone is going to come up with a reasonably good enough clone, and then their cash cow is dead.
The long term plan for Office is to turn it into the front end for a client-server document management/groupware/web publishing system. Microsoft has found out that the network is the profit center and may be ready to ceed the 'simple' word proc and spreadsheet market. --
The *only* reason that you can buy a cheap x86 motherboard is because there's a huge bolt-and-screw clone market out there. The individual build-it-yourself market is not very large.
There used to be a few bare PPC motherboards available, but they were spin-offs from the Taiwanese Mac Clone makers. When IBM dropped the "Power Personal" and Apple dropped the clones, the loose board market disappeared.
Actually, I suspect the real high end box Apache tuners are out making big consulting dollars and not reading usenet. Think I'll go post a question about processor affinity and network cards on comp.os.ms-windows.nt.networking, just to see what turns up. --
Well assuming the PR audience for the bench is Microsoft shops (because Unix/Linux shops hardly would care), I think the test makes a valid point. Linux is no silver bullet for your high end NT/IIS woes.
In fact, it's probably correct that a tuned NT/IIS system out performs an untuned Linux system. If you currently have NT/IIS, it's probably is a hellva lot smarter to try to tune what you've got before junking it for something else that you would have to go back and figure out how to tune anyways.
(This of course doesn't address IIS's instablity, which is the number one bane of IIS shops, not throughput.)
Actually, most of the performance tuning on the NT side is rather vanilla, but this is the one that struck me as being odd.
The Affinity tool allows you to bind certain processes to certain CPUs. I'm not quite sure why they'd do this with the network card, except perhaps to fix some deficency with the NT networking code when the cards are running full bore. (Which as someone pointed out, they were saturating normal ethernet.)
Note this is the exact same FUD that Microsoft used to pull when it was battling Novell for the File+Print market.
There would consistantly be benchmarks on two processor systems, even though Novell wasn't even touching the second CPU (this was the P5 era, when 2 processors were very rare.)
On the other hand, I don't think anyone has any doubt that Linux would out perform NT on a x86, where x = 5, but the type of people that see these benchmarks generally don't care about that. Benchmarks on "typical" server hardware (say a Proliant 5000 with 2 Pentium Pros and 256MB ram) is much much more interesting. --
Microsoft uses a little tree widget in some of their web pages. On IE, this renders nice and quick as an ActiveX control. Takes about an eon to load as a Java applet in Netscape. (Although, as far as I can tell Linux + Netscape + Java = lock every time.) --
Slashdot Mirror
I'll tell you -- the MS/Intel PC 99 spec specifically overrides the previous PCI spec. It's now up to the OS to allocate PCI resources, not the BIOS, although companies can chose to have a legacy mode setting.
This, and WinModems, are just the tip of the Iceburg. Microsoft is trying to turn the PC clone into the Windows computer. They've disposed of serveral critical pieces of the IBM PC AT Clone - text mode, ISA slots, and the joystick port.
It's starting to look like a PC99 will get a Windows logo, but won't even boot MS-DOS or any other OS. RMS is right on here - if Microsoft were to get away with this, it would be far worse than things like "secret" DOC file formats.
--
Just as a minor correction, Apple does compete with other Mac developers. Try to buy an iMac without paying the ClarisWorks "tax". They're used to be four or five other "works" packages for the Mac - Apple drove them under.
There's numerous other examples - Final Cut, WebObjects. I think the key is that there's no comprehensive competition like there is with Microsoft. I seriously doubt there is one Windows vendor whose #1 competition is not Microsoft themselves (well, Adobe).
--
1) Netscape doesn't yet have a componentized browser. Other companies such as Lotus, Quicken, and AOL are using the IE component, so the specs may be open enough to reverse engineer (then again, they may not.)
2) There's tons of dohickeys that plug into the Explorer shell. There's also been rip-and-replace alternative shells in the past. What's LiteStep trying to do.
3) Every Windows application manipulates the file associateions, Microsoft or not. It's a operating system design flaw (and one that KDE also seems to have). Look at Macintosh file types for a better implementation.
The big problem with Microsoft's interfaces is not that their secret (although it's conventional wisdom, is it true?), but that by that Microsoft only implements a new API along with a new product. They therefore get the time to market advantage.
As a contrary approach, look at Apple OpenDoc. They released a API, but then refused to use internally at Claris or in the Finder. As a third party vendor, the concern that an API is going to stick around for a long time is certainly valid, and when MS develops something like OLE, at least they use it themselves.
--
According to my understanding, a user can monitor what APIs are being called by an application. Someone did catch Microsoft Office using secret APIs, but that was in 1991 or something, I have not heard about a similar scandal since.
Contrary to this assertation, one reason Microsoft has gotten into corporations favor is that every product they create has a published API. Real Media has a player, Microsoft has a Media Player SDK. You probably could write a complete application suite on top of the published API for MS Office, and so on.
The only API which I understand is "secret" is the low-level WinNT kernal API. But do any MS user space applications call this API? None that I've heard of.
Many, many software companies write fully functional non-buggy Windows applications. Are they hiring ex-Microsoft employees who snuck documentation out?
Not that I'm not saying MS is not a legal monopoly and shouldn't be penalized. Just the idea that their APIs are some bastion of secrecy (like IBM's were in the 60s) doesn't necessarily hold water.
--
Every attempt at a combination Game console/Computer has not been that successful. Think Coleco Adam, Atari XE, Amiga CD, Sega Saturn plus Internet add on, Pippen, etc.
That's not to say it's a bad idea. It's just hard to design a game console (which has a shelf life of 2-5 years) that can keep up with PCs that double in speed every 18 months.
--
What's wrong with WinCE in a game console? (Other than the words "Microsoft" and "Windows" that is.)
Seems to me you could have a number of interesting convergance applications, beyond web browsing. Kind of like the Atari 400 or Commodore 64 -- primarily game machines, but you could also balance your budget or whatever.
--
Note that in most environments, the "power user" who hacking his registry or trying to install Linux on his work machine is the worst kind of user, and the kind user that generates the most support headaches.
A real power user is someone who's writing Access (or Unix equivlant) databases or HTML/Javascript and so on. It's not someone whose trying to change his system configuration or test out some cool downloads he's found. Those people are trying to get the company to subsidize (through hardware and support) their computer hobby.
Most corporations either do disallow this behavior or would if they could. Linux/Unix is the ultimate solution in this regard simply because root is restricted enough to give these folks little or no opportunity to hack on their own configuation.
Note that I'm not expressing contempt for the average user trying to do their job here - only the special case.
--
You make an extremely valid point regarding the difference between a home ("mom") system and a system that functions well in a corporate network environment. (One of the huge problems with Windows 9x is that it's trying to satisfy both these markets at the same time.)
However, in this context, Linux currently comes out behind in some respects. Linux (and Unix in general) is still very based on the host-terminal model, and therefor provides excellent remote administration capabilties, but limited capabilities to admin a group of hosts.
But in a client situation, corporations don't necessarily want better administrative tools, what's needed is less administration altogether.
Currently there's a lack of system management tools avalible, and if you did have 1000 odd Linux desktop machines, you couldn't even use something as rudimentary as Windows system policies to use default settings. In fact, as far as I can tell, you are limited to using a local account database on each machine, unless you buy Caldera's NDS module. (I could be missing NIS here.) When the #1 user complaint has traditionally been password management (too many, not syncronized, etc), this is a big problem.
So, while it may seem like a panecea to be able to telnet into Joe User's box and install or fix his word processor, what's really needed is a network-oriented system where such actions are unecessary on a user-by-user basis. X-Terminals and Java NCs have pretty much fallen by the wayside -- Maybe that's where the network admins should look, rather than trying to give the "dumb user" population Unix.
--
Don't forget that a good portion of the "Windows Tax" is actually end user technical support. The actual cost of Windows-with-no-support to OEM customers is only $25.
Any preloaded Linux setup is going to have a "RedHat tax" or a "Corel tax", just to pay for the tech support. Sure a clone shop could give you Linux and send you to the newsgroups when you have a problem, but don't expect Dell or Compaq to do that.
Admittedly, that $25 is significant with a $400 computer, but fogeys like myself still think computers cost $2000, and that $25 is probably insignificant in that scope.
--
You've hit the nail on the head right there. Various projects are making very good progress on developing a fully functional user environment for Linux, but it still seems that the system administration tools are lacking, when in fact the system admin (WinGeek) audience is probably more likely download or buy Linux than a normal user.
Someone above pointed out that there are Linux users that don't want to be Unix users, and that certainly is a valid market out there. The problem is right now, you can't be sure that the system admin tools that exist (RedHat stuff, LinuxConf) are actually working properly without breaking the "legacy" (from their point of view) unix stuff.
On a Windows or Mac system, after you've made a configuration change, 99.5% of the time you don't need to go into the configuration database to make sure that it's right. And if you do, it's rightfully labeled as a bug. Linux needs to strive for this level of functionality.
The only tools which sounds like their close are YaST (haven't tried it yet) and IBM AIX's thing.
(MasOSX may have a good solution -- A binary config database that you can dump out to standard Unix config files, which you then edit and import back.)
--
Yes, but that's not how they caught the guy. They actually got him because AOL coughed up his home phone number.
The GUID only showed that he built the virus from other, already known viruses.
--
Actually, I run the exact same configuration, except with a $3 antenna, and the picture is really damn good.
--
NTSC broadcasted over the airways will be obsolete, but your TV set will not be. You'll be able to buy and inexpensive HDTV tuner box and hook it up, much like people do with cable boxes now.
Note that most "High Definition" TV broadcast will actually be at the same resolution as normal NTSC, there'll just be more (free) channels over the airwaves.
--
Well, if you look at the average MS Word user, they probably only use 20% of the feature set.
But, if you look across a medium or large organization, you'd probably find that 95% of Word's features are actively being used.
Sure a cheaper, more lightweight, more targeted solution would solve most peoples needs most of the time, but for the more advanced features, you'd have to go with other tools. This creates an integration and deployment nightmare when you are trying to support an organizational "standard".
In short, IT departments that pay through the nose for MS Office know what they're getting. There's been cheaper/lightweight solutions many times before (Volkswriter!), but they've never flown.
--
Are you refering to IBM's decision to target OS/2 1.x to the 286?
--
This has to be one of the most retarded MS flames of all time. When the maximum amount of memory in a microcomputer was 64K, 640K was a hellva lot.
Linux -- From the guy that said that 2 GB of RAM would be enough. Just wait.
--
I recall talking to an Netscape sales rep soon after Navigator 1.1 came out. She said that Netscape wanted to give the software away for free, because it was free advertising for the server products. But big accounts kept coming and saying that they wouldn't deploy it unless they could get a support contract. So Netscape "had to" start site licescing the thing. Plausable enough to be beliveable.
There were still many many dolts that bought the shrinkwrapped Netscape box rather than download it.
--
Yeah, if you think it's bad now, just wait a couple months for the big 2000 PR bulldozer. Should knock Linux right off ZDNet.
/. legend are generally happy with NT4) are going to be understandably wary about Win2000.
Microsoft knows it's in a bit of jam with the Win2000 server upgrade. It contains an enormous number of network services that are untested and poorly understood. IT shops (who contrary to
In fact, many NT shops are still tying their networks together with NBT broadcasts and have a lot of work and brainpower to accumulate to implement Active Directory.
In short, there's a good chance that most of the installed base won't upgrade, much like most of the Novell 3.x base hiccuped when 4.x came out. This opens the door wide open for Linux, if it can act enough like an NT clone from the NOS perspective.
--
Well maybe Gates knows something not many other people do -- Microsoft Office has gotten pretty much every feature that could be possibly be devised. Pretty soon, someone is going to come up with a reasonably good enough clone, and then their cash cow is dead.
The long term plan for Office is to turn it into the front end for a client-server document management/groupware/web publishing system. Microsoft has found out that the network is the profit center and may be ready to ceed the 'simple' word proc and spreadsheet market.
--
The *only* reason that you can buy a cheap x86 motherboard is because there's a huge bolt-and-screw clone market out there. The individual build-it-yourself market is not very large.
There used to be a few bare PPC motherboards available, but they were spin-offs from the Taiwanese Mac Clone makers. When IBM dropped the "Power Personal" and Apple dropped the clones, the loose board market disappeared.
--
Not a very helpful reply, eh?
Actually, I suspect the real high end box Apache tuners are out making big consulting dollars and not reading usenet. Think I'll go post a question about processor affinity and network cards on comp.os.ms-windows.nt.networking, just to see what turns up.
--
Well assuming the PR audience for the bench is Microsoft shops (because Unix/Linux shops hardly would care), I think the test makes a valid point. Linux is no silver bullet for your high end NT/IIS woes.
In fact, it's probably correct that a tuned NT/IIS system out performs an untuned Linux system. If you currently have NT/IIS, it's probably is a hellva lot smarter to try to tune what you've got before junking it for something else that you would have to go back and figure out how to tune anyways.
(This of course doesn't address IIS's instablity, which is the number one bane of IIS shops, not throughput.)
--
Actually, most of the performance tuning on the NT side is rather vanilla, but this is the one that struck me as being odd.
The Affinity tool allows you to bind certain processes to certain CPUs. I'm not quite sure why they'd do this with the network card, except perhaps to fix some deficency with the NT networking code when the cards are running full bore. (Which as someone pointed out, they were saturating normal ethernet.)
Anyone have any other ideas why this was done?
--
Note this is the exact same FUD that Microsoft used to pull when it was battling Novell for the File+Print market.
There would consistantly be benchmarks on two processor systems, even though Novell wasn't even touching the second CPU (this was the P5 era, when 2 processors were very rare.)
On the other hand, I don't think anyone has any doubt that Linux would out perform NT on a x86, where x = 5, but the type of people that see these benchmarks generally don't care about that. Benchmarks on "typical" server hardware (say a Proliant 5000 with 2 Pentium Pros and 256MB ram) is much much more interesting.
--
Microsoft uses a little tree widget in some of their web pages. On IE, this renders nice and quick as an ActiveX control. Takes about an eon to load as a Java applet in Netscape. (Although, as far as I can tell Linux + Netscape + Java = lock every time.)
--