Oops! I'm not a Java programmer, and I didn't mean to start a flamewar about this.
My point was simply that traditional RAD client-server development isn't a good approach for the smaller and not-heavily used applications that corporations usually run to VB for. Maybe Java doesn't fit the bill as a RAD tool, but in terms of deployment and supportability (sic), it beats VB hands down. Feel free to substitute what ever other tool fits the bill. --
Yes, it would be shameful if Sun spiffed up Solaris' interface to make it more popular with users. They might actually (gasp) gain market share that way.
Basically, Sun only wants to sell expensive and esoteric server hardware. That makes them about as relevant to this discussion as IBM's S/390 division. --
VisualBasic (and friends) has been hugely influential in the move to purge Mac and Unix users from the desktop in major corporations.
However, the classic VB-style client-server front end is old news, even according to Microsoft propaganda. Get Wine working for those legacy VB apps, and work on good web or java interfaces for new client-server apps. The key product here is Mozilla, not another RAD system.
But if you insist, Borland Delphi is coming to Linux, and (as it's proponents argue) it does everything VB does, only better. --
Right on. It's importatnt to note that Microsoft has "Human Interface Guidelines" too, althought they originated at IBM in the 1980s.
On the other hand, there was a thread on the Gnome developers mailing list where they pretty much dissed the idea of having any UI standards for Gnome apps, even for little things like the placement of the Help menu. Perhaps it's too early or things are in flux, but on the surface this seems like a real bad idea. --
Look at VA-Linux..what is the difference between VA and Ma+Pa's Korner Klone Store?
VA makes some nice boxes, but face it, so do quite a few other screwdriver shops. Oh, and VA refuses to support anything but Linux, which makes them popular among Linux users, but not-so-popular among the people who buy the other 98% of Intel servers. --
Maybe, but what if, oh say M$ pumped out a perl implementatio,...
Read the linked article. Stallman addresses your concern directly. He even uses Microsoft as the example case.
Essentially he says it's not a problem because MS's Perl would still be open source, so users' could just fork it back to standard Perl, or take MS's additions and add those to standard Perl. (However, I'm not real familiar with the Artistic Licence, so perhaps MS could close the source. Stallman is of course advocating the GPL.) --
Interesting, but if someone came up with a good clean room "Java compatible" product, the value of the Java trademark would be pretty much dead. Look at Microsoft - they've had no problem selling things like "NWLink" (NetWare IPX/SPX) and Visual "J" (Java).
As someone else pointed out, The Open Group's UNIX trademark is pretty much worthless when the most accessible versions of Unix don't even meet the criteria.
Sun probably knows that the clock is ticking and there's only so much time left that they can throw around the weight of the Java trademark. If IBM or Microsoft comes up with a clean room Java, I would expect Sun to go running back to ECMA or ISO or whoever. --
You are correct, "UNIX(tm)" is a specification, no longer a specific product, so in theory a Linux or BSD OS could be certified.
However, pragmatically, the UNIX spec requires non-free software like CDE and Motif which most Linux OSes wouldn't use. In addition, I've heard that GNU isn't very interested in making their stuff 100% compliant with UNIX/POSIX specs. --
Perhaps Joe Blow Quake Player doesn't care about compatibility, but for any decent MIS department, compatibility and standardization are the foremost concern, because that's the single largest thing you can do to reduce costs.
This is a big reason that Windows and MS Office has been forced onto virtually every corporate desktop. By going with one vendor's end-to-end standard, you radically reduce compatibility problems, and lower costs significantly in the short term. The reason they aren't worried about Unix users' ability to read MS Word docs or IE web sites or Lotus Notes e-mail is because the easiest solution they have is to get rid of Unix.
Now, I'm worried that I might be misread - I'm not advocating a MS or any heterogenous solution by any means. In fact, I tend to agree with the pedantic system analysts that in the long term the solution is open XML document formats and open network protocols. However, that wheel hasn't been entirely invented yet, so in the short term, I sympathize with people's pragmatic concerns. --
Many argue that the only reason RMS wants to free Java is so he can fork it. That's rubbish
In fact his point is exactly the opposite. He argues that if (say) Microsoft forked GNU-Java, users could either fork it right back, or take Microsoft's improvements and roll them into other implementations. The result is more compatibility and more features.
Although the folks at javalobby probably don't want to hear it, Java is going to fork eventually. Pretty soon IBM and MS and everyone else is just going to give up on Sun and sell their own version of Sun-"compatible" HotBrownLiquid. If the compatible versions are open source, everyone is better off than if there are proprietary forks. --
Furthermore, it's suprising how many Slashdot readers seem to be missing the entire point of reverse-engineering Unix -- Software is supposed to be source compatible with any Unix system.
Indeed, leaving all of the politics aside, one of the biggest real-world freedoms granted by an Open Source Unix platform is that you are not tied to specific hardware, a specific vendor, or even a specific OS kernel. It shouldn't matter if you are running any given Linux or any given BSD OS, because the cost of migrating between them is relatively low.
Instead, around here, it's always Linux versus BSD versus Solaris versus the world. Which I find odd, because considering the tiny market share Unix has on the desktop and low-end servers, it would seem that Unix as a platform should be advocated, not necessarily any specific implementation. The message to the NT-using rest of the world is getting drowned out in the noise of the fraternal politics. --
The average friggin' moron will be just as friggin' moronic as he is today
It's a lost cause, pal. People are always going to be morons. But that doesn't mean you should deny them access to certain tools. In fact the goal should be to make computer systems more accesible to the morons (without dumbing them down for the non-morons).
Look at this way: From an engineering standpoint, developing a complex, hard and expensive to administer Unix system is boring. It's been done a hundred times, and will be done a hundred more times.
The Linux vendors, on the other hand, are attempting something that's never successfully been done before in the history of consumer computing -- they are trying to build a system which meets the needs of both the Dumb User population and the Unix Wizard User population. If they pull it off, it will be one of the greatest engineering feats of all time. --
If there was something wrong with text installers, people would be scared away from Windows 2000, which does it's initial setup in text mode also.
The bad taste in my mouth comes from RedHat's (and Mandrake's) old text mode install program. The color scheme is hard to read and the keyboard navigation is confusing. Not critical problems, but it does make it seem harder than it really is. And if you are going to rewrite the installer to correct these problems, you might as invest in a GUI version. --
Yes, those of us who have been on the Internet know about reliable retailers such as Amazon or FatBrain, and generally shop there. But, in the last few months, the number of online retailers have exploded. I watched a football game on TV, and there were far more *.com commercials than there were car commercials!
I'm willing to bet a good number of these companies are on a make-money-fast business plan, and have invested far more in television commercials than in information systems. But on the Internet, how do you know? Anyone with a big ad budget and a flashy site design looks just as competent as Amazon. The local news had a piece about some company that would take on-line orders, print them out, and then phone in the order to a middleman. Yeah, that's e-commerce all right.
So, considering the large number of companies showing up late and underprepared for the gold rush, and the large number of newbie Internet users that don't have loyalties to existing sites, I would imagine that more goes wrong than you and I would see from ordering from the old standbys.
(My only bad experience was a series of repeated attempts to buy a 21" monitor from buy.com. Each attempted transaction ended in the black hole of a MS-ODBC ASP error. Furthermore, their phone representative couldn't take a telephone order, and couldn't confirm if my order had even been placed or not. I had to call my credit card to make sure that I wasn't charged. Bottom line -- it was a pain in the ass, and there's $1100 that buy.com didn't make.) --
Right, the loader partition needs to be accessible, but once it is the NT loader can read a SCSI driver from a FAT partition and then boot from any partition on the system, even if it's (say) the fourth partition on the fifth hard disk. With LILO, you might be able to boot from/dev/sdf4, and you might not - it's depends on your BIOS.
Note that LILO works fine in most situations, but I would consider it kludger than the NT loader. AFAICT, it basically just remembers the physical location of the kernel on disk and jumps there and starts running. The NT loader cheats by being able to read NTFS volumes, which makes it more reliable in strange situations. --
I've seen alot of FUD from Microsoft, but in this case I would take this "gem of 'support information'" for exactly that.
You might want to read this post, check your paranoia at the door, and keep your powder dry. Microsoft has better weapons at it's disposal than the support database. --
One thing that I've heard is that Microsoft phone-in tech support uses a similar Knowledge Base to the one that's online. If something is documented in the KB, the caller will be charged for the call.
I can imagine that they've had a few "Duhh - I installed Linux on my computer cuz I was trying to be elite and now I can't install NT to use PowerPoint!"
So there's no conspiricy here, other than to put the information in the hands of the idiot first level phone support people and get another 75 bucks or whatever out of callers, --
Considering Microsoft has offered instructions on how to convert OS/2* and Novell systems for years, Linux users should see this as a compliment! They're finally in the big league.
* DOS 6.0 included a text file which stated something like "Congratulations on your choice to upgrade OS/2 to MS-DOS 6.0!" Some upgrade. --
Even Windows 3.1 didn't have the sort of monitor problems that plague X. Sure, but you're usually stuck with 60 Hz SVGA and a very limited choice of resolutions. And there's no way to get it to run on more exotic hardware like fixed frequency Workstation monitors.
It's been a long time, but installing a video driver for Windows 3.1 was really damn difficult, and usually involved editing your SYSTEM.INI file by hand. Most people just left the setting at SVGA 800x600x16 and lived with it. --
As others have pointed out, Windows users are usually getting 60 frames per second...
This statement is flat-out incorrect. On a Windows 9x system with modern hardware, the refresh rate will be close to the maximum the hardware supports at any given resolution. Of course, it's still Windows, so YMMV.
that's a matter of every hardware vendor (like Matrox and Diamond and SIII etc) writing drivers and configurations for Windows.
Well, as mentioned above, Windows has a standard mechanism for monitor manufacturer to tell the OS about supported resolutions. (INF files) On the other hand, every different Linux OS distribution seems to have it's own X configuration program with different datafile formats. Before demanding manufacturer support for Linux in this department, the vendors need to get their act together and define a standard way that this information can be published. --
, Linux & GNU (GNU/Linux anyone?) are becoming the 'Standard Unix'.
Your point only proves the pointlessness of your argument. If a PC workstation is running a (Linux/*BSD/Solaris/UnixWare) kernel and has the GNU toolset and (KDE/Gnome/WindowMaker/etc) installed, it is going to be virtually indistingishble from a user's standpoint from any other PC Unix system with a similar setup. The differences will only manifest themselves when you are doing system administration, but that's not what you are talking about.
So stop the pointless advocacy of Unix kernel A over Unix kernel B, because by-in-large users don't care about Unix kernel services. You are better off spending your time advocating *nix itself over other alternatives (especially because it's become a non-entity on the desktop, partially due to such dickering such as yours.) --
Along the lines of IBM's stuggle with OS/2, don't forget the original goal of Windows NT was to be an ultraportable operating system, and NT4 shipped with support for 4 platforms. Unfortunately, in a 'work-alike' situation, the platform with the cost and mindshare advantage won 98% of NT's market and NT5 is now shipping with x86 support only. (However, I've heard rumors that WinCE is stripped-down NT kernel and that does run on several CPUs.)
The moral of IBM's and MS's story is that maintaining a multi-platform commercial OS is a money loser. (Sun keeps making the poorly supported Solaris x86, apparently as an intellectual exercise.)
Just another case where Linux breaks all the normal rules. --
No, I don't think that claim is ridiculous. The person who buys networking equipment is most likely not the "end user" (meaning the person who is actually using the NIC/Hub). It's the person who is paid to worry about such things.
(You and I may have home networks for our personal use, but that's an exception to the rule.) --
Re:Itanium (Merced) will be relevant right away...
on
News on Pentium IV
·
· Score: 1
I sure could be wrong, but my understanding is that's not correct -- the presence of 16-bit asm in the OS/2 kernel was one of the reasons Microsoft rewrote that part for NT, and IBM wrote a new microkernel for the PowerPC version. Of course, now that Warp is on version 5, that problem could have been solved long ago.
Also, I think OS/2 2.0 actually shipped in 1990. (On the same day as Windows 3.0?) If it was 1992, they would have only been one year ahead of MS's 32-bit OS.
Anyway, my original point was that the i386 shipped in 1997, and even to this day most users aren't running a 32-bit clean OS. The whole 32-bit transition has gone marvelously bad -- let's hope the 64-bit transition goes better. --
Re:Itanium (Merced) will be relevant right away...
on
News on Pentium IV
·
· Score: 1
OS/2 1.x was all 16-bit because IBM targeted at it's 286-based PS/2 line of computers. Even to this day, there's still bits of 16-bit code in OS/2.
Not that that's necessarily bad, but I'm not sure if it's the model we want to hold up for the 32 to 64 bit transition. --
Slashdot Mirror
Oops! I'm not a Java programmer, and I didn't mean to start a flamewar about this.
My point was simply that traditional RAD client-server development isn't a good approach for the smaller and not-heavily used applications that corporations usually run to VB for. Maybe Java doesn't fit the bill as a RAD tool, but in terms of deployment and supportability (sic), it beats VB hands down. Feel free to substitute what ever other tool fits the bill.
--
Yes, it would be shameful if Sun spiffed up Solaris' interface to make it more popular with users. They might actually (gasp) gain market share that way.
Basically, Sun only wants to sell expensive and esoteric server hardware. That makes them about as relevant to this discussion as IBM's S/390 division.
--
VisualBasic (and friends) has been hugely influential in the move to purge Mac and Unix users from the desktop in major corporations.
However, the classic VB-style client-server front end is old news, even according to Microsoft propaganda. Get Wine working for those legacy VB apps, and work on good web or java interfaces for new client-server apps. The key product here is Mozilla, not another RAD system.
But if you insist, Borland Delphi is coming to Linux, and (as it's proponents argue) it does everything VB does, only better.
--
Right on. It's importatnt to note that Microsoft has "Human Interface Guidelines" too, althought they originated at IBM in the 1980s.
On the other hand, there was a thread on the Gnome developers mailing list where they pretty much dissed the idea of having any UI standards for Gnome apps, even for little things like the placement of the Help menu. Perhaps it's too early or things are in flux, but on the surface this seems like a real bad idea.
--
Look at VA-Linux..what is the difference between VA and Ma+Pa's Korner Klone Store?
VA makes some nice boxes, but face it, so do quite a few other screwdriver shops. Oh, and VA refuses to support anything but Linux, which makes them popular among Linux users, but not-so-popular among the people who buy the other 98% of Intel servers.
--
Maybe, but what if, oh say M$ pumped out a perl implementatio,...
Read the linked article. Stallman addresses your concern directly. He even uses Microsoft as the example case.
Essentially he says it's not a problem because MS's Perl would still be open source, so users' could just fork it back to standard Perl, or take MS's additions and add those to standard Perl. (However, I'm not real familiar with the Artistic Licence, so perhaps MS could close the source. Stallman is of course advocating the GPL.)
--
Interesting, but if someone came up with a good clean room "Java compatible" product, the value of the Java trademark would be pretty much dead. Look at Microsoft - they've had no problem selling things like "NWLink" (NetWare IPX/SPX) and Visual "J" (Java).
As someone else pointed out, The Open Group's UNIX trademark is pretty much worthless when the most accessible versions of Unix don't even meet the criteria.
Sun probably knows that the clock is ticking and there's only so much time left that they can throw around the weight of the Java trademark. If IBM or Microsoft comes up with a clean room Java, I would expect Sun to go running back to ECMA or ISO or whoever.
--
You are correct, "UNIX(tm)" is a specification, no longer a specific product, so in theory a Linux or BSD OS could be certified.
However, pragmatically, the UNIX spec requires non-free software like CDE and Motif which most Linux OSes wouldn't use. In addition, I've heard that GNU isn't very interested in making their stuff 100% compliant with UNIX/POSIX specs.
--
Perhaps Joe Blow Quake Player doesn't care about compatibility, but for any decent MIS department, compatibility and standardization are the foremost concern, because that's the single largest thing you can do to reduce costs.
This is a big reason that Windows and MS Office has been forced onto virtually every corporate desktop. By going with one vendor's end-to-end standard, you radically reduce compatibility problems, and lower costs significantly in the short term. The reason they aren't worried about Unix users' ability to read MS Word docs or IE web sites or Lotus Notes e-mail is because the easiest solution they have is to get rid of Unix.
Now, I'm worried that I might be misread - I'm not advocating a MS or any heterogenous solution by any means. In fact, I tend to agree with the pedantic system analysts that in the long term the solution is open XML document formats and open network protocols. However, that wheel hasn't been entirely invented yet, so in the short term, I sympathize with people's pragmatic concerns.
--
Many argue that the only reason RMS wants to free Java is so he can fork it. That's rubbish
In fact his point is exactly the opposite. He argues that if (say) Microsoft forked GNU-Java, users could either fork it right back, or take Microsoft's improvements and roll them into other implementations. The result is more compatibility and more features.
Although the folks at javalobby probably don't want to hear it, Java is going to fork eventually. Pretty soon IBM and MS and everyone else is just going to give up on Sun and sell their own version of Sun-"compatible" HotBrownLiquid. If the compatible versions are open source, everyone is better off than if there are proprietary forks.
--
Furthermore, it's suprising how many Slashdot readers seem to be missing the entire point of reverse-engineering Unix -- Software is supposed to be source compatible with any Unix system.
Indeed, leaving all of the politics aside, one of the biggest real-world freedoms granted by an Open Source Unix platform is that you are not tied to specific hardware, a specific vendor, or even a specific OS kernel. It shouldn't matter if you are running any given Linux or any given BSD OS, because the cost of migrating between them is relatively low.
Instead, around here, it's always Linux versus BSD versus Solaris versus the world. Which I find odd, because considering the tiny market share Unix has on the desktop and low-end servers, it would seem that Unix as a platform should be advocated, not necessarily any specific implementation. The message to the NT-using rest of the world is getting drowned out in the noise of the fraternal politics.
--
The average friggin' moron will be just as friggin' moronic as he is today
It's a lost cause, pal. People are always going to be morons. But that doesn't mean you should deny them access to certain tools. In fact the goal should be to make computer systems more accesible to the morons (without dumbing them down for the non-morons).
Look at this way: From an engineering standpoint, developing a complex, hard and expensive to administer Unix system is boring. It's been done a hundred times, and will be done a hundred more times.
The Linux vendors, on the other hand, are attempting something that's never successfully been done before in the history of consumer computing -- they are trying to build a system which meets the needs of both the Dumb User population and the Unix Wizard User population. If they pull it off, it will be one of the greatest engineering feats of all time.
--
If there was something wrong with text installers, people would be scared away from Windows 2000, which does it's initial setup in text mode also.
The bad taste in my mouth comes from RedHat's (and Mandrake's) old text mode install program. The color scheme is hard to read and the keyboard navigation is confusing. Not critical problems, but it does make it seem harder than it really is. And if you are going to rewrite the installer to correct these problems, you might as invest in a GUI version.
--
Yes, those of us who have been on the Internet know about reliable retailers such as Amazon or FatBrain, and generally shop there. But, in the last few months, the number of online retailers have exploded. I watched a football game on TV, and there were far more *.com commercials than there were car commercials!
I'm willing to bet a good number of these companies are on a make-money-fast business plan, and have invested far more in television commercials than in information systems. But on the Internet, how do you know? Anyone with a big ad budget and a flashy site design looks just as competent as Amazon. The local news had a piece about some company that would take on-line orders, print them out, and then phone in the order to a middleman. Yeah, that's e-commerce all right.
So, considering the large number of companies showing up late and underprepared for the gold rush, and the large number of newbie Internet users that don't have loyalties to existing sites, I would imagine that more goes wrong than you and I would see from ordering from the old standbys.
(My only bad experience was a series of repeated attempts to buy a 21" monitor from buy.com. Each attempted transaction ended in the black hole of a MS-ODBC ASP error. Furthermore, their phone representative couldn't take a telephone order, and couldn't confirm if my order had even been placed or not. I had to call my credit card to make sure that I wasn't charged. Bottom line -- it was a pain in the ass, and there's $1100 that buy.com didn't make.)
--
Right, the loader partition needs to be accessible, but once it is the NT loader can read a SCSI driver from a FAT partition and then boot from any partition on the system, even if it's (say) the fourth partition on the fifth hard disk. With LILO, you might be able to boot from /dev/sdf4, and you might not - it's depends on your BIOS.
Note that LILO works fine in most situations, but I would consider it kludger than the NT loader. AFAICT, it basically just remembers the physical location of the kernel on disk and jumps there and starts running. The NT loader cheats by being able to read NTFS volumes, which makes it more reliable in strange situations.
--
I've seen alot of FUD from Microsoft, but in this case I would take this "gem of 'support information'" for exactly that.
You might want to read this post, check your paranoia at the door, and keep your powder dry. Microsoft has better weapons at it's disposal than the support database.
--
One thing that I've heard is that Microsoft phone-in tech support uses a similar Knowledge Base to the one that's online. If something is documented in the KB, the caller will be charged for the call.
I can imagine that they've had a few "Duhh - I installed Linux on my computer cuz I was trying to be elite and now I can't install NT to use PowerPoint!"
So there's no conspiricy here, other than to put the information in the hands of the idiot first level phone support people and get another 75 bucks or whatever out of callers,
--
Considering Microsoft has offered instructions on how to convert OS/2* and Novell systems for years, Linux users should see this as a compliment! They're finally in the big league.
* DOS 6.0 included a text file which stated something like "Congratulations on your choice to upgrade OS/2 to MS-DOS 6.0!" Some upgrade.
--
Even Windows 3.1 didn't have the sort of monitor problems that plague X.
Sure, but you're usually stuck with 60 Hz SVGA and a very limited choice of resolutions. And there's no way to get it to run on more exotic hardware like fixed frequency Workstation monitors.
It's been a long time, but installing a video driver for Windows 3.1 was really damn difficult, and usually involved editing your SYSTEM.INI file by hand. Most people just left the setting at SVGA 800x600x16 and lived with it.
--
As others have pointed out, Windows users are usually getting 60 frames per second ...
This statement is flat-out incorrect. On a Windows 9x system with modern hardware, the refresh rate will be close to the maximum the hardware supports at any given resolution. Of course, it's still Windows, so YMMV.
that's a matter of every hardware vendor (like Matrox and Diamond and SIII etc) writing drivers and configurations for Windows.
Well, as mentioned above, Windows has a standard mechanism for monitor manufacturer to tell the OS about supported resolutions. (INF files) On the other hand, every different Linux OS distribution seems to have it's own X configuration program with different datafile formats. Before demanding manufacturer support for Linux in this department, the vendors need to get their act together and define a standard way that this information can be published.
--
, Linux & GNU (GNU/Linux anyone?) are becoming the 'Standard Unix'.
Your point only proves the pointlessness of your argument. If a PC workstation is running a (Linux/*BSD/Solaris/UnixWare) kernel and has the GNU toolset and (KDE/Gnome/WindowMaker/etc) installed, it is going to be virtually indistingishble from a user's standpoint from any other PC Unix system with a similar setup. The differences will only manifest themselves when you are doing system administration, but that's not what you are talking about.
So stop the pointless advocacy of Unix kernel A over Unix kernel B, because by-in-large users don't care about Unix kernel services. You are better off spending your time advocating *nix itself over other alternatives (especially because it's become a non-entity on the desktop, partially due to such dickering such as yours.)
--
Along the lines of IBM's stuggle with OS/2, don't forget the original goal of Windows NT was to be an ultraportable operating system, and NT4 shipped with support for 4 platforms. Unfortunately, in a 'work-alike' situation, the platform with the cost and mindshare advantage won 98% of NT's market and NT5 is now shipping with x86 support only. (However, I've heard rumors that WinCE is stripped-down NT kernel and that does run on several CPUs.)
The moral of IBM's and MS's story is that maintaining a multi-platform commercial OS is a money loser. (Sun keeps making the poorly supported Solaris x86, apparently as an intellectual exercise.)
Just another case where Linux breaks all the normal rules.
--
No, I don't think that claim is ridiculous. The person who buys networking equipment is most likely not the "end user" (meaning the person who is actually using the NIC/Hub). It's the person who is paid to worry about such things.
(You and I may have home networks for our personal use, but that's an exception to the rule.)
--
I sure could be wrong, but my understanding is that's not correct -- the presence of 16-bit asm in the OS/2 kernel was one of the reasons Microsoft rewrote that part for NT, and IBM wrote a new microkernel for the PowerPC version. Of course, now that Warp is on version 5, that problem could have been solved long ago.
Also, I think OS/2 2.0 actually shipped in 1990. (On the same day as Windows 3.0?) If it was 1992, they would have only been one year ahead of MS's 32-bit OS.
Anyway, my original point was that the i386 shipped in 1997, and even to this day most users aren't running a 32-bit clean OS. The whole 32-bit transition has gone marvelously bad -- let's hope the 64-bit transition goes better.
--
OS/2 1.x was all 16-bit because IBM targeted at it's 286-based PS/2 line of computers. Even to this day, there's still bits of 16-bit code in OS/2.
Not that that's necessarily bad, but I'm not sure if it's the model we want to hold up for the 32 to 64 bit transition.
--