With the exception of fat binaries, what you describe was the strategy DEC had for NT-on-Alpha (with the FX32 emulator). It didn't work for DEC/Compaq - why would it work for AMD?
IIRC, Microsoft wanted something like $200 Million from Motorola for a NT5 port to PPC. Considering AMD is somewhat cash strapped, they'd do better spending that money on x86 R+D, as the x86 platform is going to be viable for at least the next 5-10 years. --
That sucks. But what's stopping people from doing NAT with Windows NT, if they have driver support for this card? (Besides the obvious slashdot commentary about WinNT.)
--
Re:OS/2 doesn't matter (its deader than the Amiga
on
G4 vs. Athlon Review
·
· Score: 1
My point was that that there's a very small market for a low-end PC that can't run 'legacy' or old OSes. If you wanted such a thing, you might as well just buy an Alpha. --
Don't worry, I understand the concepts here, and as an occasional Mac user, I can't wait until OS X comes out. My primary platform is Win NT, so I can attest to the benefits of a reasonably sound implementation of modern OS concepts (although the Linux folks will probably disagree!), and just deal with rebooting if I need DOS or Win 9x games.
My point is simply that in the real world implementation matters more than buzzword compliance. Despite it's modern feature set, an app crash in Windows 9x is just as likely to lead to a reboot as one in MacOS. Multitasking is smoother under Windows 9x, but most Mac programs are written to allow long jobs to be run in the background, so the users hardly notice the difference.
The real problem here is MS's tendancy to kludge unstable bits onto the OS (such as ActiveDesktop and that horrid Win98 ACPI stuff) before fixing what they've got. Despite Apple's utter failure to get a modern OS out the door, they've been doing a hellava lot of bugfixing over the years, and ye olde MacOS is actually quite stable despite it's architectural faults.
{From a marketing standpoint, Windows 9x is just bizzare -- with WinNT sitting (rotting?) on the shelf for so many years, Microsoft has insisted on dumping engineering into the old DOS/Win platform. Backwards compatbility *can't* be that important to the user base. It's like Apple coming out with MacOS X and then continuing with 5 years of development of the old MacOS. I just don't understand it.} --
I always thought it was an interesting coincidence that you (?and Jobs) did some work for Atari before you started Apple, and that Atari was able to ship a personal computer shortly after the Apple ][, especially since the Apple and the Atari were fairly similar machines.
I'm curious -- what sort of work did you do for Atari? (I've heard rumors that you designed the Breakout arcade game.) Did you pitch your computer to the management at Atari? If so, what was their reaction? Did they ever give you any legal hassles over including "Breakout" with the Apple? --
Yes, Windows 9x proves that you can be fully buzzword compliant, have excellent backwards compatibility (even on the driver level!), and still have an unstable POS OS.
Not that MacOS is a paradigm of stability, but I think it stands up pretty well to Win 9x, even while failing to provide modern OS services and also having piss-poor backwards compatibilty for anything older than 3-4 years. Either the Mac community is so rabid to be blind to these faults, or in the real world these concerns are vastly overblown.
I'm not getting into a flamewar either -- only making the point that from the average user's point of view, the unmodern nature of the MacOS is a non-issue (and in fact has probably made some interesting audio/video applications that demand 'bare metal' access possible). From a programmers standpoint, I can see how one would think the MacOS sucks. --
Well, you hear rumors about AMD's 64-bit "K8" chip, which would have a unique ISA. However, the only OS that would run on this chip would be Linux, unless AMD forked Microsoft some cash to port NT. Since the Linux market isn't nearly big enough to support a unique ISA aimed at the low-end, the K8 talk sounds more like anti-Intel marketing FUD than anything for real.
OS support is the biggest reason the Apple-style emulator approach wouldn't work in the PC world. DOS still does matter, as does OS/2, NetWare, BeOS, and old versions of Windows. Since it's unlikely that a emulator would ever be added to these OSes, you'd be cutting yourself off from the 'legacy' market, and essentially end up like NT/Alpha. --
And then you will see "MacOS X Server 2.1" and MacOS X Client 3.0", so the "X" is really becoming part of the product name (and will probably be pronounced "echhs" and not "ten".) --
There seems to be an aversion (even among managers in MIS departments) to software that has a very high minor number. It sorta implies that the vendor has patched the software to heck and can't get the new version out the door.
Compare "Lotus Notes R4.67c" versus "MS Outlook 98" -- Even though the Lotus scheme is honest major.minor, the Microsoft version implies more progress (even though in reality there's been almost no progress, it really should be called "Exchange Client 2.1").
So, I can understand why Sun made the switch from 2.x to X. They aren't just trying communicate with the guys in the back room anymore, they need to market to CIOs and other dunderheads that are going to veto upgrades unless it seems really worthwhile.
One word of wisdom from the depths of DLL Hell is that all files installed by a particular version of Windows have been stamped with the exact date and same version number. Thus, all files in Win95 OSR2.5 will appear newer than the ones in OSR2.1, even if they haven't been updated.
This sort of scheme can have marginal benefits in a support situation, but it also can be a detriment too, especially with third party installers that mung their date stamps so that they can overwrite you system DLLs. (Thinking too much about Windows -- now my head is really starting to hurt.) --
I can fairly say that both the 97 and 2000 versions improved the product to the point that most consumers would want the upgrade.
I'm curious what in Office 2000 was worth the upgrade, in your opinion. Except for the HTML/XML export and a zanier paperclip, Office 2000 has a nearly identical feature set to Office 97.
Microsoft sends my company this stuff for free, so I use it. But I can't imagine paying for it and feeling like it was worth it. --
Agreed. I don't know what's worse -- the DCOM approach of using a mystery port between 34567 and 65432, or tunnelling RPC calls over plain 'ol port 80 and hoping the Unix hippie firewall admin doesn't notice. Either way you are opening yourself to remote code execution that may not be expected. --
If you're going to use Windows NT, you should probably keep that firewall in place between those Windows service ports and the rest of the world. Microsoft loves to add services and open ports to your computer when you're not looking. And it's probably not going to be the IP stack, it'll probably be some goofy listening service, like anonymous share enumeration or something. Or maybe remote access to NetDDE. Or some authentication protocol that doesn't like large Netbios fields. Or possibly even some undocumented functionality in the named pipe filesystem used for RPC. Who knows. Personally, I'm not going to wait around to find out.
"Firewall your NT systems!" -- This bit of advice has been widely known by experienced NT admins for many years -- some existing vulnerabilities having been documented back in the OS/2 LanMan era in the late 80s. Like early Unix network protocols, the product was designed for a mostly-trusted LAN environment, and this design philosophy has been continued with even fairly new add-ons like MS Transaction Server.
Unfortunately, with the huge growth of NT as a platform, shifty or incorrect Microsoft documentation, an education program (MCSE) that completely neglects these issues, and a generally ignorant group of low-end administrators, there is a huge number of unprotected NT systems running on the Internet. (Compare this to Unix, where there exists a broad understanding of Internet security issues, and a healthy community skepticism of security claims.) As time, home broadband, and Windows 2000 goes on, I would expect that the number of unsecured hosts is going to out number the firewalled ones.
Considering the underlying culture, I doubt an "Internet Worm II - This time it's NT!" would lead to anything more than a cosmetic fix. Unfortunately, Microsoft is probably going to have to redesign the control mechanisms of the numerous RPC services that run on NT and create a nice GUI with a big "Internet (Secure) Mode" checkbox. A security blanket, but it's going to do nothing to educate the administrators or engender a culture of security consciousness, and exploits will continue.
Just as the original Internet Worm didn't shift the tide away from Unix and towards VMS, I really doubt these issues will affect marketshare seriously. Only, as the number of specialized Internet hosts grows, Unix's compartmentalized, peer-reviewed approach is going to continue to win over Microsoft's poorly understood philosophy of integrated RPC services. --
(The supposedly non-standard) Kerberos authentication in NT5 is going to break Samba bad. The old NTLM protocols are still supported, but as soon as a shop goes all Windows 2000, those should be turned off, and it's going to be hard to put a Samba server in the corner and not have anyone know it. --
It was actually a good solution for Apple. First of all, they have a transparent 68K emulator, so unlike other solutions, the user wasn't aware what code was native and what was emulated.
Second, the Mac FAT binary wasn't necessarily two entire binaries in one file. You could have mix of PPC and 68K code and it run seemlessly. This allowed vendors to port the speed critical stuff first, and get around to converting the rest of the 68K stuff whenever they felt like it. (A good chunk of legacy API functions in the MacOS is still 68K.)
Third, with a simple drag-n-drop utility, PPC users could automatically weed out the unnecessary 68K code resources, making the binary thin again. However, the disk space wastage from fat binaries was usually pretty minimal.
Obviously, it's a dumb solution for Unix systems, especially open source ones, but it worked well for the Mac. --
Considering that new $500 computer has more than enough juice to run Windows 2000, I think your post is largely wishful thinking. Maybe in the world of cast-off 486s and systems with 32MB of RAM, but that doesn't describe any current corporate desktop I'm aware of. Microsoft knows who they're marketing the product at, and those people have plenty of hardware juice (usually underutilized by running Win 9x.) --
Re:BFD. Just disable the power saving crap.
on
Laptop Pentium IIIs
·
· Score: 1
Nope, certain CPUs (PowerPC?) have a variable clock hardwired on all the time. Wouldn't be shocked if Intel started doing this too. --
I'm suspicious at the clock speed claims of these new chips. I have Dell Mobile Pentium II-450 at work, and the BIOS Setup contains a warning that setting the CPU to Full Speed may create an unstable system. This leads one to think that the CPU slows itself down depending on temperature, and may in fact never be running at rated speed. --
Look at all the items promised by MS but will not be delivered when Windows2000 shows up.
Like what? Load-balancing component server support? NetWare file system support? 24-way CPU support? It's not like Linux has these things either, and it's doubtful that shops will miss them.
Windows 2000 is two years late, but the major pieces are there - ActiveDirectory, dynamic DNS, remote installation support, plug-n-play. The real question isn't the missing features, but the fact that the average MS shop is going to be scared of the complexity of this stuff. --
Yup, recall the "OS/2 Hysteria" of 1994-95. Consumer magazines started to print articles, lots of nerds started duel booting, flamewars-er-advocacy abound on the net, and it all came to a crashing halt when Windows 95 and NT 4.0 finally made it out the door in somewhat good shape.
Not that there is anything wrong with either Linux or OS/2, but neither of them was really designed to meet the frustrations of the mainstream computing user. However, I think Linux will ride the "crash" a little better than OS/2 did. For one, the price is still right. For two, the momentium of the computing industry now is right towards web services, which happens to be Linux's strong point. (Where OS/2's strong points were host gateways or an object desktop or something more ambigious.) --
Having more than two stories is against the law (zoning codes) in most of Silicon Valley.
You are correct that life is becoming unbearable here -- rampant land speculation and thousands of paper millionares have created the worst bubble economy in the country. The corporate solution is to import foreigners on work visas and stay put in SV because of the massive amount of liquid capital being thrown around. However, pretty soon the labor issues are going to hit the fan, and then, while I have no hope for San Jose, maybe San Francisco at least will become a nice place to live again. --
Slashdot Mirror
With the exception of fat binaries, what you describe was the strategy DEC had for NT-on-Alpha (with the FX32 emulator). It didn't work for DEC/Compaq - why would it work for AMD?
IIRC, Microsoft wanted something like $200 Million from Motorola for a NT5 port to PPC. Considering AMD is somewhat cash strapped, they'd do better spending that money on x86 R+D, as the x86 platform is going to be viable for at least the next 5-10 years.
--
That sucks. But what's stopping people from doing NAT with Windows NT, if they have driver support for this card? (Besides the obvious slashdot commentary about WinNT.)
--
My point was that that there's a very small market for a low-end PC that can't run 'legacy' or old OSes. If you wanted such a thing, you might as well just buy an Alpha.
--
Don't worry, I understand the concepts here, and as an occasional Mac user, I can't wait until OS X comes out. My primary platform is Win NT, so I can attest to the benefits of a reasonably sound implementation of modern OS concepts (although the Linux folks will probably disagree!), and just deal with rebooting if I need DOS or Win 9x games.
My point is simply that in the real world implementation matters more than buzzword compliance. Despite it's modern feature set, an app crash in Windows 9x is just as likely to lead to a reboot as one in MacOS. Multitasking is smoother under Windows 9x, but most Mac programs are written to allow long jobs to be run in the background, so the users hardly notice the difference.
The real problem here is MS's tendancy to kludge unstable bits onto the OS (such as ActiveDesktop and that horrid Win98 ACPI stuff) before fixing what they've got. Despite Apple's utter failure to get a modern OS out the door, they've been doing a hellava lot of bugfixing over the years, and ye olde MacOS is actually quite stable despite it's architectural faults.
{From a marketing standpoint, Windows 9x is just bizzare -- with WinNT sitting (rotting?) on the shelf for so many years, Microsoft has insisted on dumping engineering into the old DOS/Win platform. Backwards compatbility *can't* be that important to the user base. It's like Apple coming out with MacOS X and then continuing with 5 years of development of the old MacOS. I just don't understand it.}
--
I always thought it was an interesting coincidence that you (?and Jobs) did some work for Atari before you started Apple, and that Atari was able to ship a personal computer shortly after the Apple ][, especially since the Apple and the Atari were fairly similar machines.
I'm curious -- what sort of work did you do for Atari? (I've heard rumors that you designed the Breakout arcade game.) Did you pitch your computer to the management at Atari? If so, what was their reaction? Did they ever give you any legal hassles over including "Breakout" with the Apple?
--
Yes, Windows 9x proves that you can be fully buzzword compliant, have excellent backwards compatibility (even on the driver level!), and still have an unstable POS OS.
Not that MacOS is a paradigm of stability, but I think it stands up pretty well to Win 9x, even while failing to provide modern OS services and also having piss-poor backwards compatibilty for anything older than 3-4 years. Either the Mac community is so rabid to be blind to these faults, or in the real world these concerns are vastly overblown.
I'm not getting into a flamewar either -- only making the point that from the average user's point of view, the unmodern nature of the MacOS is a non-issue (and in fact has probably made some interesting audio/video applications that demand 'bare metal' access possible). From a programmers standpoint, I can see how one would think the MacOS sucks.
--
Well, you hear rumors about AMD's 64-bit "K8" chip, which would have a unique ISA. However, the only OS that would run on this chip would be Linux, unless AMD forked Microsoft some cash to port NT. Since the Linux market isn't nearly big enough to support a unique ISA aimed at the low-end, the K8 talk sounds more like anti-Intel marketing FUD than anything for real.
OS support is the biggest reason the Apple-style emulator approach wouldn't work in the PC world. DOS still does matter, as does OS/2, NetWare, BeOS, and old versions of Windows. Since it's unlikely that a emulator would ever be added to these OSes, you'd be cutting yourself off from the 'legacy' market, and essentially end up like NT/Alpha.
--
Yup, I know. However, what percentage of Office 2000 shops have implemented this? My guess it's pretty low.
Ask Prof. Clippy "What's new in Office 2000?" and he will give you a very thin list of new features.
--
Solaris A? (Unless Sun is on an octal system.)
--
And then you will see "MacOS X Server 2.1" and MacOS X Client 3.0", so the "X" is really becoming part of the product name (and will probably be pronounced "echhs" and not "ten".)
--
There seems to be an aversion (even among managers in MIS departments) to software that has a very high minor number. It sorta implies that the vendor has patched the software to heck and can't get the new version out the door.
Compare "Lotus Notes R4.67c" versus "MS Outlook 98" -- Even though the Lotus scheme is honest major.minor, the Microsoft version implies more progress (even though in reality there's been almost no progress, it really should be called "Exchange Client 2.1").
So, I can understand why Sun made the switch from 2.x to X. They aren't just trying communicate with the guys in the back room anymore, they need to market to CIOs and other dunderheads that are going to veto upgrades unless it seems really worthwhile.
--
One word of wisdom from the depths of DLL Hell is that all files installed by a particular version of Windows have been stamped with the exact date and same version number. Thus, all files in Win95 OSR2.5 will appear newer than the ones in OSR2.1, even if they haven't been updated.
This sort of scheme can have marginal benefits in a support situation, but it also can be a detriment too, especially with third party installers that mung their date stamps so that they can overwrite you system DLLs. (Thinking too much about Windows -- now my head is really starting to hurt.)
--
I can fairly say that both the 97 and 2000 versions improved the product to the point that most consumers would want the upgrade.
I'm curious what in Office 2000 was worth the upgrade, in your opinion. Except for the HTML/XML export and a zanier paperclip, Office 2000 has a nearly identical feature set to Office 97.
Microsoft sends my company this stuff for free, so I use it. But I can't imagine paying for it and feeling like it was worth it.
--
Agreed. I don't know what's worse -- the DCOM approach of using a mystery port between 34567 and 65432, or tunnelling RPC calls over plain 'ol port 80 and hoping the Unix hippie firewall admin doesn't notice. Either way you are opening yourself to remote code execution that may not be expected.
--
If you're going to use Windows NT, you should probably keep that firewall in place between those Windows service ports and the rest of the world. Microsoft loves to add services and open ports to your computer when you're not looking. And it's probably not going to be the IP stack, it'll probably be some goofy listening service, like anonymous share enumeration or something. Or maybe remote access to NetDDE. Or some authentication protocol that doesn't like large Netbios fields. Or possibly even some undocumented functionality in the named pipe filesystem used for RPC. Who knows. Personally, I'm not going to wait around to find out.
"Firewall your NT systems!" -- This bit of advice has been widely known by experienced NT admins for many years -- some existing vulnerabilities having been documented back in the OS/2 LanMan era in the late 80s. Like early Unix network protocols, the product was designed for a mostly-trusted LAN environment, and this design philosophy has been continued with even fairly new add-ons like MS Transaction Server.
Unfortunately, with the huge growth of NT as a platform, shifty or incorrect Microsoft documentation, an education program (MCSE) that completely neglects these issues, and a generally ignorant group of low-end administrators, there is a huge number of unprotected NT systems running on the Internet. (Compare this to Unix, where there exists a broad understanding of Internet security issues, and a healthy community skepticism of security claims.) As time, home broadband, and Windows 2000 goes on, I would expect that the number of unsecured hosts is going to out number the firewalled ones.
Considering the underlying culture, I doubt an "Internet Worm II - This time it's NT!" would lead to anything more than a cosmetic fix. Unfortunately, Microsoft is probably going to have to redesign the control mechanisms of the numerous RPC services that run on NT and create a nice GUI with a big "Internet (Secure) Mode" checkbox. A security blanket, but it's going to do nothing to educate the administrators or engender a culture of security consciousness, and exploits will continue.
Just as the original Internet Worm didn't shift the tide away from Unix and towards VMS, I really doubt these issues will affect marketshare seriously. Only, as the number of specialized Internet hosts grows, Unix's compartmentalized, peer-reviewed approach is going to continue to win over Microsoft's poorly understood philosophy of integrated RPC services.
--
(The supposedly non-standard) Kerberos authentication in NT5 is going to break Samba bad. The old NTLM protocols are still supported, but as soon as a shop goes all Windows 2000, those should be turned off, and it's going to be hard to put a Samba server in the corner and not have anyone know it.
--
It was actually a good solution for Apple. First of all, they have a transparent 68K emulator, so unlike other solutions, the user wasn't aware what code was native and what was emulated.
Second, the Mac FAT binary wasn't necessarily two entire binaries in one file. You could have mix of PPC and 68K code and it run seemlessly. This allowed vendors to port the speed critical stuff first, and get around to converting the rest of the 68K stuff whenever they felt like it. (A good chunk of legacy API functions in the MacOS is still 68K.)
Third, with a simple drag-n-drop utility, PPC users could automatically weed out the unnecessary 68K code resources, making the binary thin again. However, the disk space wastage from fat binaries was usually pretty minimal.
Obviously, it's a dumb solution for Unix systems, especially open source ones, but it worked well for the Mac.
--
Considering that new $500 computer has more than enough juice to run Windows 2000, I think your post is largely wishful thinking. Maybe in the world of cast-off 486s and systems with 32MB of RAM, but that doesn't describe any current corporate desktop I'm aware of. Microsoft knows who they're marketing the product at, and those people have plenty of hardware juice (usually underutilized by running Win 9x.)
--
Nope, certain CPUs (PowerPC?) have a variable clock hardwired on all the time. Wouldn't be shocked if Intel started doing this too.
--
Yes, apparently Linux doesn't like variable speed CPUs. See here for more info. Apparently, it's a pretty serious problem.
--
I'm suspicious at the clock speed claims of these new chips. I have Dell Mobile Pentium II-450 at work, and the BIOS Setup contains a warning that setting the CPU to Full Speed may create an unstable system. This leads one to think that the CPU slows itself down depending on temperature, and may in fact never be running at rated speed.
--
Look at all the items promised by MS but will not be delivered when Windows2000 shows up.
Like what? Load-balancing component server support? NetWare file system support? 24-way CPU support? It's not like Linux has these things either, and it's doubtful that shops will miss them.
Windows 2000 is two years late, but the major pieces are there - ActiveDirectory, dynamic DNS, remote installation support, plug-n-play. The real question isn't the missing features, but the fact that the average MS shop is going to be scared of the complexity of this stuff.
--
Yup, recall the "OS/2 Hysteria" of 1994-95. Consumer magazines started to print articles, lots of nerds started duel booting, flamewars-er-advocacy abound on the net, and it all came to a crashing halt when Windows 95 and NT 4.0 finally made it out the door in somewhat good shape.
Not that there is anything wrong with either Linux or OS/2, but neither of them was really designed to meet the frustrations of the mainstream computing user. However, I think Linux will ride the "crash" a little better than OS/2 did. For one, the price is still right. For two, the momentium of the computing industry now is right towards web services, which happens to be Linux's strong point. (Where OS/2's strong points were host gateways or an object desktop or something more ambigious.)
--
You can always move to the Bay Area, where someone with your experience will get $100/easily, but still be just as poor!
--
Having more than two stories is against the law (zoning codes) in most of Silicon Valley.
You are correct that life is becoming unbearable here -- rampant land speculation and thousands of paper millionares have created the worst bubble economy in the country. The corporate solution is to import foreigners on work visas and stay put in SV because of the massive amount of liquid capital being thrown around. However, pretty soon the labor issues are going to hit the fan, and then, while I have no hope for San Jose, maybe San Francisco at least will become a nice place to live again.
--