IBM puts up with, and makes an enormous amount of money of Windows NT/2000 in services. Much more then they make off of Linux right now.
All in all, they'd rather have RedHat get their small cut rather than Microsoft, just as long as they get to move their hardware and services. Plus Linux+Netfinity gets them at a portion of Sun's market that AIX+RS/6000 doesn't. --
Glad I got the attention of an OpenServer expert of sorts...
The problem I see is that if creating a SCO System Services Emulation Layer was easy, SCO themselves would have done it for UnixWare. It can't be any easier on Linux, and frankly, I can't expect Caldera to care that much about OpenServer customers except to milk them for whatever possible.
I guess I'm skeptical about SCO's great "VAR Channel". As far as I can tell, it's a bunch of vertical market apps that are in legacy mode to the extent that the vendors can't/won't port them to anything up-to-date. Which is fine as a 'legacy' revenue channel, but totally a non-issue for the further expansion of Linux or Unixware. What Caldera got in UnixWare is an underpromoted, full-fledged OS which they can push to Unix-friendly x86-friendly Linux shops. --
Are you trying to say "Desktop Integration is Bad", or are you saying that "The Unix architecture is so convoluted that nobody can do Desktop Integration correctly"?
Whatever, but Integrated Desktop users on (say) MacOS or BeOS don't have to trace through 4 layers of script to get it working. The Magic happens to be great in fact for people on those platforms.
What you are saying is that the stack of cards has gotten so high that it falls down randomly. Bad Thing, but you realize what Miguel of Gnome is saying about the suckiness lack of policy in the Unix infrastructure. --
That goes both ways too. Caldera apparently feels that UnixWare plus the Linuxy GNU/KDE/Gnome/XFree stuff will appeal to Linux users who may have grown out of the Linux kernel. (Clustering, lots of CPUs, etc.)
This is a risky strategy because now it's unknown if Caldera is fully supportive of $49 Linux, or are they trying to upsell you to $Thousands UnixWare. Hopefully they'll cut the UnixWare price.
Of course, when the inherited the VAR channel, they also inherited SCO UNIX (OpenServer). My understanding was that SCO wasn't having much luck getting the customer base off of OpenServer and onto UnixWare. Maybe Linux will be a better sell, but having a 'legacy' customer base can be a real pain in the ass. (Ask Compaq, who bought DEC for the services business, but ended up becoming DEC after finding that VMS and DEC Unix couldn't so easily be swept under the rug.)
128KB was an obscene amount of memory in the late 1970s for a personal device. (For example, the Atari 800 computer had a maximum of 48KB and usually shipped with less in the early years.)
I happen to know that Microsoft has a secret project to going port Win32 to a clean room, portable, SMP non-GPL kernel. The goal is to run Win32 executables in a stable, network-aware environment. In addition to Win32 the kernel will support UNIX and OS/2 personalities.
I'd tell you the name of this project, but then I'd have to kill you! --
The "integration" wasn't even attempted. Novell went from "UnixWare is the SuperNOS of the future" to "Netware + NDS is the future (with WinNT as your application server)" in a very short amount of time.
Maybe the customer base didn't want UnixWare. On the other hand, they weren't so enthusastic about NetWare 4.x either. Anyway, now Novell has been spending years doing things like adding TCP/IP and turning NW into a web/database server to appease it's 15% marketshare of true believers. Too bad that a few years ago they owned all of that and more in UNIX. --
Right, they planned Plug'n'Play and Power Mangement for NT4, and they just plain missed the feature cut. Fine. That's no excuse for putting the technology on the shelf for four years.
When IE4 and the Option Pack (IIS + Transaction Server) rolled out in 98 or so, there was a lot of compliaints even from the Microsoftie fan base that they really should have shipped NT4.5.
In my view, the only reason they didn't was that they were lying through their teeth about the difficulties they were having with Active Directory, and consequentally the NT5 ship date kept shifiting from 1997 to 1998 to 1999 to finally early 2000.
Basically, workstation users like me got screwed, and server admins got screwed (By NT4+Option Pack's convoluted install process), all in the name of marketing to super huge corporations with thousands of nodes and NDS envy. NT4.5 could have been the "universal OS" -- it certainly would have been a step up from 95/98. --
Yup, the Windows NT project started out as a project to rewrite the OS/2 kernel. This was before the IBM-Microsoft divorce, and both companies ended up owning the technologies that they co-developed (OS/2 v1 and Windows v3.0). Consequentially, there's a bunch of OS/2-isms in NT (CMD.EXE, NTFS, the entire SMB networking system, etc.) --
I can say "device drivers", and after three years of running NT 4.0 as my primary desktop OS, I never found a non-USB peripheral for which I could not say "device driver".
Sure, there's some consumer stuff out there with no NT drivers (such as parallel port stuff), but it's easy enough to avoid. And I can understand their point of view -- as long as Microsoft hasn't committed to NT as consumer platform, why should they?
(PS: Gaming drivers are an issue, I agree. However, even then NT4 + Matrox drivers were good enough for my modest needs: Age of Empires, SimCity, Quake, etc.) --
I can say "device drivers", and after three years of running NT 4.0 as my primary desktop OS, I never found a non-USB peripheral for which I could not say "device driver".
Sure, there's some consumer stuff out there with no NT drivers (such as parallel port stuff), but it's easy enough to avoid. And I can understand their point of view -- as long as Microsoft hasn't committed to NT as consumer platform, why should they?
(PS: Gaming drivers are an issue, I agree. However, even then NT4 + Matrox drivers were good enough for my modest needs: Age of Empires, SimCity, Quake, etc.) --
On my box, WOW doesn't load until a 16-bit app is launched.
And WOW is a safer approach. It translates Win16 calls into Win32 calls. That means that illegal Win16 calls can be caught by the WOW layer and the app can be crashed.
Compare this to 9x's approach. The core OS still contains the whole Windows 3.1 core. The 32-bit stuff is kludged on above, aside, and underneath it with thunking layers. App makes an illegal Win16 call and the OS either crashes or becomes unpredictable. Furthermore, you still may be suseptible to Win16 bugs in the core OS which date back 10 years or more. --
To clarify, Windows 95 had a shadow CONFIG.SYS/AUTOEXEC.BAT, which loaded such things as HIMEM.SYS and SETVER and of course WIN. Since 95 wouldn't boot without this functionality, it's reasonable to think that it's still there in ME. --
The entire Win32s for Windows 3.1 and Windows 4.x series (95, 98, 98SE, ME) has existed for one purpose: get all applications replaced by Win32 versions and migrate everyone to NT.
The original intent of Win32s and Windows 95 might have been to be a stepping stone to NT, but Microsoft has been strangely really lax about the execution. They've almost purpously kept NT out of the mainstream market by keeping things like Plug'n'Play, USB support, and DirectX on the shelf for 3 to 4 years after the 9x folks got it.
I wonder if this is a situation similar to Apple in the 1980s, where the "Apple II forever" people kept the lineup on the market for years past the point where it was competitive. It's almost as if iinternal forces inside Microsoft are conspiring to keep the 95/98/ME line going, and when upper management looks at the revenue figures, they can't argue.
Don't forget, these guys have a monopoly on the desktop. If they wanted the world to run the NT kernel, they could get us there. Instead, they want to treat Windows 2000 as an upsell and continually pedddling crap like Windows ME. The sell of Windows 95 was "Just like NT, except with backwards compatibility". But now the sell of ME seems to be "Just like 98, except without the backwards compability". Bizarre. --
Once, in the early days of IE 4.0, I went to a page and got a nice dialog box asking me if I wanted to install "IE Destroyer" and whether I should "always trust ScriptKiddies?". By Microsoft's standards, that was a trusted, verified application. (I was pretty happy that I'd monked with the default settings at that point.) --
Those are only the bugs they admitted to. Furthermore, the never seem to mention the 'features' they've slipped in along the way either. (For an enjoyable experience take a look at the Lotus Notes fix lists -- each point release's list is about twice as long as Microsoft's, and there's a new release every couple months. But, at least it's honest.)
Of course, Microsoft only admits to some of the bugs because of people like you who will point at the list and say See How Long It Is, I Told You So! Of course, you probably use software from a vendor that can't even be bothered to print fix lists! --
Re:I would like to see TCP/IP work
on
Plex86 Runs DOS
·
· Score: 1
Microsoft's drivers work pretty well, although there's no software to go with them other than an SMB client. (They're on every NT4 CD in the clients directory. Rumor had it MS used to have a DOS telnet and FTP program, but I've looked and can't find it. )
Arachne was bought and renamed to something else by Caldera's DR-DOS side and more development was done. You might want to try it again - it seemed fairly stable to me, although slow. It sits on NetWare ODI drivers, IIRC. --
My understanding is that Microsoft uses the term "MAPI" for two different things:
1) The method by which Windows programs (Word, for example) figures out how to send mail through your mail program. This has been supplimented somewhat by the mailto: way, which is done via URL protocol registration (IE integration). Most Windows mail programs support this.
2) The API used to implement client-server services for only Outlook/Windows Messaging/Exchange Client. This is how Outlook talks to Exchange, for example, as well as supporting Internet Mail (and I think Compuserve and MSN at one time.) The Outlook-to-Exchange wire protocol is often referred to as "MAPI", and that's what I was saying, but I think that's techncially incorrect. My understanding is that the Exchange MAPI service uses standard Windows RPC to talk to the server. --
Since Lotus has shipped server software for Linux, that means their C API kit is available. With a little work, it shouldn't be too difficult to talk to Domino/Notes with the native protocol. --
Really, the only thing Outlook/Exchange brings to the "corporate communications" market is a pretty user interface, and a nicely functional calendar module.
Beneth the skin, the mail functionality is only so-so, the search engine sucks eggs, it's virtually impossible to effectively customize Outlook or develop your own applications, the security is braindead, and the server is unstable and doesn't scale worth crap.
Unfortunately, because of the lack of calendaring software options (NetWare-specific Groupwise, "legacy" iPlanet/Netscape, and expensive/complex/ugly Lotus, and nothing from the open protocol world), lots of shops end up *having* to buy Exchange.
But the only way you can cost-justify any of these things is to effectively use the groupware functionality, which is so braindead in Ex/Out as to almost be worthless. Microsoft has lots of big plans for this market, but to date, they've really delivered squat.
Anyway, lots of people use Exchange, but it's just one of those things that nobody is totally happy about, except for Microsoft who gets an easy sale in many shops. --
I thought Go got swollowed by AT+T, who then put them in a few commercials ("You Will!") and then after an executive decision that AT+T really didn't want to be in the computer industry, got shut down.
Anyway, hardly a fair test of Pen computing, much like the Apple Newton (which probably was introduced two years too early).
(BTW, There used to be a WWW page lauding Go's OS as the greatest OO environment ever, beating even NeXTStep, but I can't seem to find it now.) --
Slashdot Mirror
IBM puts up with, and makes an enormous amount of money of Windows NT/2000 in services. Much more then they make off of Linux right now.
All in all, they'd rather have RedHat get their small cut rather than Microsoft, just as long as they get to move their hardware and services. Plus Linux+Netfinity gets them at a portion of Sun's market that AIX+RS/6000 doesn't.
--
Glad I got the attention of an OpenServer expert of sorts...
The problem I see is that if creating a SCO System Services Emulation Layer was easy, SCO themselves would have done it for UnixWare. It can't be any easier on Linux, and frankly, I can't expect Caldera to care that much about OpenServer customers except to milk them for whatever possible.
I guess I'm skeptical about SCO's great "VAR Channel". As far as I can tell, it's a bunch of vertical market apps that are in legacy mode to the extent that the vendors can't/won't port them to anything up-to-date. Which is fine as a 'legacy' revenue channel, but totally a non-issue for the further expansion of Linux or Unixware. What Caldera got in UnixWare is an underpromoted, full-fledged OS which they can push to Unix-friendly x86-friendly Linux shops.
--
Are you trying to say "Desktop Integration is Bad", or are you saying that "The Unix architecture is so convoluted that nobody can do Desktop Integration correctly"?
Whatever, but Integrated Desktop users on (say) MacOS or BeOS don't have to trace through 4 layers of script to get it working. The Magic happens to be great in fact for people on those platforms.
What you are saying is that the stack of cards has gotten so high that it falls down randomly. Bad Thing, but you realize what Miguel of Gnome is saying about the suckiness lack of policy in the Unix infrastructure.
--
That goes both ways too. Caldera apparently feels that UnixWare plus the Linuxy GNU/KDE/Gnome/XFree stuff will appeal to Linux users who may have grown out of the Linux kernel. (Clustering, lots of CPUs, etc.)
This is a risky strategy because now it's unknown if Caldera is fully supportive of $49 Linux, or are they trying to upsell you to $Thousands UnixWare. Hopefully they'll cut the UnixWare price.
Of course, when the inherited the VAR channel, they also inherited SCO UNIX (OpenServer). My understanding was that SCO wasn't having much luck getting the customer base off of OpenServer and onto UnixWare. Maybe Linux will be a better sell, but having a 'legacy' customer base can be a real pain in the ass. (Ask Compaq, who bought DEC for the services business, but ended up becoming DEC after finding that VMS and DEC Unix couldn't so easily be swept under the rug.)
--
Like he said: "No, Really".
128KB was an obscene amount of memory in the late 1970s for a personal device. (For example, the Atari 800 computer had a maximum of 48KB and usually shipped with less in the early years.)
But thanks anyway for your attempt as wiseassery.
--
Check out www.interix.com and your local OS2.EXE binary. Windows Ten shipped seven years ago.
Hopefully I've convinced one person how silly this MS Linux postuation is. (grin!)
--
It's still there. They just removed the documentation.
--
I happen to know that Microsoft has a secret project to going port Win32 to a clean room, portable, SMP non-GPL kernel. The goal is to run Win32 executables in a stable, network-aware environment. In addition to Win32 the kernel will support UNIX and OS/2 personalities.
I'd tell you the name of this project, but then I'd have to kill you!
--
This integration was such an abysmal failure
The "integration" wasn't even attempted. Novell went from "UnixWare is the SuperNOS of the future" to "Netware + NDS is the future (with WinNT as your application server)" in a very short amount of time.
Maybe the customer base didn't want UnixWare. On the other hand, they weren't so enthusastic about NetWare 4.x either. Anyway, now Novell has been spending years doing things like adding TCP/IP and turning NW into a web/database server to appease it's 15% marketshare of true believers. Too bad that a few years ago they owned all of that and more in UNIX.
--
Right, they planned Plug'n'Play and Power Mangement for NT4, and they just plain missed the feature cut. Fine. That's no excuse for putting the technology on the shelf for four years.
When IE4 and the Option Pack (IIS + Transaction Server) rolled out in 98 or so, there was a lot of compliaints even from the Microsoftie fan base that they really should have shipped NT4.5.
In my view, the only reason they didn't was that they were lying through their teeth about the difficulties they were having with Active Directory, and consequentally the NT5 ship date kept shifiting from 1997 to 1998 to 1999 to finally early 2000.
Basically, workstation users like me got screwed, and server admins got screwed (By NT4+Option Pack's convoluted install process), all in the name of marketing to super huge corporations with thousands of nodes and NDS envy. NT4.5 could have been the "universal OS" -- it certainly would have been a step up from 95/98.
--
Yup, the Windows NT project started out as a project to rewrite the OS/2 kernel. This was before the IBM-Microsoft divorce, and both companies ended up owning the technologies that they co-developed (OS/2 v1 and Windows v3.0). Consequentially, there's a bunch of OS/2-isms in NT (CMD.EXE, NTFS, the entire SMB networking system, etc.)
--
I can say "device drivers", and after three years of running NT 4.0 as my primary desktop OS, I never found a non-USB peripheral for which I could not say "device driver".
Sure, there's some consumer stuff out there with no NT drivers (such as parallel port stuff), but it's easy enough to avoid. And I can understand their point of view -- as long as Microsoft hasn't committed to NT as consumer platform, why should they?
(PS: Gaming drivers are an issue, I agree. However, even then NT4 + Matrox drivers were good enough for my modest needs: Age of Empires, SimCity, Quake, etc.)
--
I can say "device drivers", and after three years of running NT 4.0 as my primary desktop OS, I never found a non-USB peripheral for which I could not say "device driver".
Sure, there's some consumer stuff out there with no NT drivers (such as parallel port stuff), but it's easy enough to avoid. And I can understand their point of view -- as long as Microsoft hasn't committed to NT as consumer platform, why should they?
(PS: Gaming drivers are an issue, I agree. However, even then NT4 + Matrox drivers were good enough for my modest needs: Age of Empires, SimCity, Quake, etc.)
--
On my box, WOW doesn't load until a 16-bit app is launched.
And WOW is a safer approach. It translates Win16 calls into Win32 calls. That means that illegal Win16 calls can be caught by the WOW layer and the app can be crashed.
Compare this to 9x's approach. The core OS still contains the whole Windows 3.1 core. The 32-bit stuff is kludged on above, aside, and underneath it with thunking layers. App makes an illegal Win16 call and the OS either crashes or becomes unpredictable. Furthermore, you still may be suseptible to Win16 bugs in the core OS which date back 10 years or more.
--
To clarify, Windows 95 had a shadow CONFIG.SYS/AUTOEXEC.BAT, which loaded such things as HIMEM.SYS and SETVER and of course WIN. Since 95 wouldn't boot without this functionality, it's reasonable to think that it's still there in ME.
--
The entire Win32s for Windows 3.1 and Windows 4.x series (95, 98, 98SE, ME) has existed for one purpose: get all applications replaced by Win32 versions and migrate everyone to NT.
The original intent of Win32s and Windows 95 might have been to be a stepping stone to NT, but Microsoft has been strangely really lax about the execution. They've almost purpously kept NT out of the mainstream market by keeping things like Plug'n'Play, USB support, and DirectX on the shelf for 3 to 4 years after the 9x folks got it.
I wonder if this is a situation similar to Apple in the 1980s, where the "Apple II forever" people kept the lineup on the market for years past the point where it was competitive. It's almost as if iinternal forces inside Microsoft are conspiring to keep the 95/98/ME line going, and when upper management looks at the revenue figures, they can't argue.
Don't forget, these guys have a monopoly on the desktop. If they wanted the world to run the NT kernel, they could get us there. Instead, they want to treat Windows 2000 as an upsell and continually pedddling crap like Windows ME. The sell of Windows 95 was "Just like NT, except with backwards compatibility". But now the sell of ME seems to be "Just like 98, except without the backwards compability". Bizarre.
--
NT on RISC used SoftPC and SoftWindows 16-bit (by Connectix? It's changed hands a few times.) Later on DEC wrote a 32-bit emulator.
NT on Intel's "DOS VM" came out of OS/2 v1, I think. It emulates some of DOS 5.0's system functions, but doesn't do any CPU emulation, of course.
--
Once, in the early days of IE 4.0, I went to a page and got a nice dialog box asking me if I wanted to install "IE Destroyer" and whether I should "always trust ScriptKiddies?". By Microsoft's standards, that was a trusted, verified application. (I was pretty happy that I'd monked with the default settings at that point.)
--
Those are only the bugs they admitted to. Furthermore, the never seem to mention the 'features' they've slipped in along the way either. (For an enjoyable experience take a look at the Lotus Notes fix lists -- each point release's list is about twice as long as Microsoft's, and there's a new release every couple months. But, at least it's honest.)
Of course, Microsoft only admits to some of the bugs because of people like you who will point at the list and say See How Long It Is, I Told You So! Of course, you probably use software from a vendor that can't even be bothered to print fix lists!
--
Microsoft's drivers work pretty well, although there's no software to go with them other than an SMB client. (They're on every NT4 CD in the clients directory. Rumor had it MS used to have a DOS telnet and FTP program, but I've looked and can't find it. )
Arachne was bought and renamed to something else by Caldera's DR-DOS side and more development was done. You might want to try it again - it seemed fairly stable to me, although slow. It sits on NetWare ODI drivers, IIRC.
--
My understanding is that Microsoft uses the term "MAPI" for two different things:
1) The method by which Windows programs (Word, for example) figures out how to send mail through your mail program. This has been supplimented somewhat by the mailto: way, which is done via URL protocol registration (IE integration). Most Windows mail programs support this.
2) The API used to implement client-server services for only Outlook/Windows Messaging/Exchange Client. This is how Outlook talks to Exchange, for example, as well as supporting Internet Mail (and I think Compuserve and MSN at one time.) The Outlook-to-Exchange wire protocol is often referred to as "MAPI", and that's what I was saying, but I think that's techncially incorrect. My understanding is that the Exchange MAPI service uses standard Windows RPC to talk to the server.
--
Since Lotus has shipped server software for Linux, that means their C API kit is available. With a little work, it shouldn't be too difficult to talk to Domino/Notes with the native protocol.
--
Really, the only thing Outlook/Exchange brings to the "corporate communications" market is a pretty user interface, and a nicely functional calendar module.
Beneth the skin, the mail functionality is only so-so, the search engine sucks eggs, it's virtually impossible to effectively customize Outlook or develop your own applications, the security is braindead, and the server is unstable and doesn't scale worth crap.
Unfortunately, because of the lack of calendaring software options (NetWare-specific Groupwise, "legacy" iPlanet/Netscape, and expensive/complex/ugly Lotus, and nothing from the open protocol world), lots of shops end up *having* to buy Exchange.
But the only way you can cost-justify any of these things is to effectively use the groupware functionality, which is so braindead in Ex/Out as to almost be worthless. Microsoft has lots of big plans for this market, but to date, they've really delivered squat.
Anyway, lots of people use Exchange, but it's just one of those things that nobody is totally happy about, except for Microsoft who gets an easy sale in many shops.
--
I really don't think that MAPI is just IMAP + NTLM. There's just far too many Exchange-isms in MAPI that don't have any parallel in IMAP.
BTW, if someone has coded an open version of NTLM, it sure would be nice if they submitted it to Mozilla.
--
I thought Go got swollowed by AT+T, who then put them in a few commercials ("You Will!") and then after an executive decision that AT+T really didn't want to be in the computer industry, got shut down.
Anyway, hardly a fair test of Pen computing, much like the Apple Newton (which probably was introduced two years too early).
(BTW, There used to be a WWW page lauding Go's OS as the greatest OO environment ever, beating even NeXTStep, but I can't seem to find it now.)
--