"and if you were a developer chances are you'd rank Eclipse very high in your list of indispensable tools."
I guess you mean if I were an Open Source developer. I hate to break it to you, but this developer along with the majority of all developers don't use Eclipse at all.
But my point was that IBM makes money in so many ways that it's hardly a test case for the business viability of Open Source.
Red Hat is a better example but probably still hasn't made a profit overall. That's pretty amazing considering that 90% of their development was done for free.
I'll be convinced when I see companies make a profit creating orignal software that is completely open and without any proprietary extensions.
"Yes a bunch of guys running Autocad might have an issue and it probably wouldn't make sense to switch them. But are they really the users still running Win2K? "
Actually, I have archietect friend who's company was running Windows 95 and the latest version of AutoCad a year ago. They've upgraded to Windows XP now, but the point is that if they stayed with Windows 95 so long, it's plausible that other firms using AutoCad might be sticking with Win2k.
"But to suggest none of the products previously list will be Open Sourced is one heck of a prediction."
IBM refuses to open source some products like Visual Test that they no longer sell, so it's one heck of a prediction that any big money maker will be open sourced as long as it's still making money.
"Why, because if you don't you are missing out on developer mind share."
You're assuming that all developers care about a product being F/OSS. At least today, the vast majority of development is being done on proprietary products.
"I avoid proprietary software for moral reasons. I expect corporate support for my ideology based on sensible business practices. "
So if MS avoids F/OSS for sensible business practices you shouldn't complain I guess.
"You can bet however that Big Blue will Open source thier big money spinners as soon as it does make sense (read will make them more money)."
You can also bet that the day it makes sense to IBM will never come. I'll bet that IBM has plans for proprietary products that they haven't even started working on yet.
In fact, here is a great question to ask IBM executives. "Will you commit to making all new software products open source by the end of this decade?"
First, you're assuming that those customers don't like Windows XP as opposed to simply chosing to stay with a version that currently meets their needs.
Second, you're assuming that if they finally feel forced to upgrade because support is ending they'll choose an alternative that isn't supported at all, doesn't run their current applications and requires a steep learning curve to use.
If they like Windows 2000 then they're certainly going to like Windows XP more than Linux.
The spin on this story seems to be that MS will be hurt because their customers aren't upgrading to Win XP.
The other fact that this story reveals is that many MS customers are so happy with Win 2K that they don't want to change. That inertia is far more damaging to the prospect of Linux on the desktop than it is to MS's bottom line.
"WebSphere Application Server is the market leading application server, I'm pretty sure. IBM eclipsed (no pun intended) BEA some time ago."
I guess the Application Server market is defined as Internet applications running on non-Apache web servers.
Overall, it still looks like IBM isn't an undisputed market leader in any category and a lot of their success seems to revolve around acquisitions rather than internal development.
I'm well aware of IBM's purchase of Rational since it resulted in the abandonment of Rational Visual Test customers (including myself and my clients). But buying software companies is not an indication that you know how to develop software. Quite the opposite I would argue.
"As I pointed out very early on in this, er, discussion, BIOS calls would be problematic, but even so, a file system with decent (and reserved) permissions would have gone a long way to discouraging a lot of bad things."
There were three common ways to program the PC, use the OS, use the BIOS and do your own thing. Very few successful commerical applications used the OS method exclusively because it was the slowest way.
If your intent all along was to say that DOS was unstable and insecure with respect to how it enforced the rules for applications that only used the OS, then I partially agree with you. Assuming they had sufficient resources they could have made the file system more secure as you suggest.
Note however that the key stability issue still couldn't be solved by the OS because application bugs could still have pointers pointing to any memory location and the 8088 didn't have any mechanism for restricting access.
"The fact that DOS on a 386 was just as promiscuous as DOS on a 8088 just proves my point that security was not a design consideration for MS, even when the hardware tools were available and waiting to be used."
This is a bit of goalpost moving on your part. The issue was whether the original design of DOS was less secure than it could have been given the limitations of the orginal PC. The 386 wasn't even in the design phase at that time.
"However, you and the other guy using your UID are having your differences, so I'll leave you two to finish the debate."
Cute but not exactly a graceful response to an apology.
What software product does IBM make that is a market leader? Most of their software isn't even ranked #2.
So it makes sense for IBM try other approaches to improve their market position. It would be a mistake for the market leaders to drop their more effective processes and adopt IBM's revised techniques until IBM can prove their value in the market.
I didn't read your post carefully enough before I made my last comment. Perhaps you do understand how to do things without involving the OS, so my comment about your skills was out of line. It's just your conclusions that are I believe are wrong.
It's true that DOS on the 386 was probably no more secure than DOS on the 8088 because MS didn't rewrite DOS to take advantage of the 386 capabilities. But that doesn't prove that one can write a secure and stable OS for the 8088.
On 8088-based machine the only way you could insure that self-modifying code was not possible would be to have the code in read-only memory. On the other hand, on a 386 the OS could setup the memory management in such a way that an application that is loaded doesn't have write access to its own code space.
Likewise, as long as interrupt vectors are stored in RAM they can be hijacked on the 8088 regardless of the OS.
I suspect (based on your description of your experience) that you already know much of what I explained here but don't want to admit that the original DOS wasn't that bad given the limitations of the orginal PC and that the primary techniques used to get around DOS could have been used against any OS in an 8088-based machine.
"And which was more stable and capable than (quick and dirty) MS-DOS, backing up my claim."
So your claim is that CP/M-86 which could have been implemented for the PC but wasn't, would have been more stable and capable than MS-DOS which was. Sounds speculative to me. So explain to me exactly how CP/M-86 would have been more stable than MS-DOS given the fact that the 8088 would still allow any application to access any memory and write to any hardware device without permission of the OS.
It wasn't my intention to misrepresent your argument or put words in your mouth but you do keep trying to bring this argument forward to the present with things like: "Which has been the design principle that MS has been stuck with ever since." I was merely anticipating an argument you I thought you were hinting at. If that was unfair, I'm sorry.
"Huh? You could overwrite system files, write self-modifying code, build code on the fly in the data segment and run it from there, steal interrupt vectors, modify the MBR, the list is nearly endless."
First of all, you were talking about users, not programmers.
If you really wrote code for the orginal PC and think you couldn't do those things without cooperation from an OS, you weren't very good at it.
As I asked before, please explain what technique a 8088-based OS could do to prevent (for example) an application program from overwriting a system file when the hardware allows writes to the disk directly without the OS being involved.
"My alternative: turn on, type URL of site. (GASP! Sounds like a "hard to use" CLI!!)"
If you're going to turn this into a GUI vs. CLI argument then by all means use the console browser. You want to go back to the previous page, type "b a c k". Gee I wonder what a web site would look like with pictures?
"The MS policy, until very recently, has always been to let a user modify anything."
There was very little that a user could modify in DOS. But my main point was that basic security and stability must be supported by the hardware; which the OS can take advantage of or not. The 8088 doesn't have such features, so applications can ignore the OS altogether and access hardware any way they please. This would be just as true for Unix as it was for DOS.
"Although BIOS calls would be a problem, any group of CS students could write an 808x OS more secure than MS-DOS."
If by this you mean that they could write a more secure OS for the original PC configuration (which we've been talking about), I doubt it. Most CS students would find it a major challenge to write any OS for that platform and for good reason.
You'd have to write most if not all of it in assembly. I suspect that most CS students spend very little time writing assembly language programs because so much of industry work is done in high level languages. That's not to say that they're not smart enough, they just don't have the training for it.
"Again, don't lecture me on MS security until they allow all users to do security updates on their systems."
I'm not lecturing you on MS security in general, but I am pointing out some misconceptions that you have about early PCs.
OK, so CP/M-86 could have been the PC's OS. I never claimed that DOS was the only possible PC OS IBM could have used, just that the limitations of the orginal PC precluded using an OS that we would consider stable and secure by today's standards.
Please explain to me how you implement privilege levels in an OS on a processor that doesn't support it. All programs running on an 8088 have full access to the entire address space and there's nothing an OS can do about it.
What reliable secure contemporary of DOS and early Windows ran on a PC? In any case, I wasn't providing an excuse for Windows based on its roots, I was responding to the claim that MS had historically always done a poor job. The fact was that more capable OS's were MIA at the time simply because they couldn't be implemented on the platform or wouldn't be any more stable if they had.
"That's been the problem with MS operating systems and the MS philosophy all along."
Well, you can't implement a secure and stable OS like Linux on a 8088 with 16K or RAM, so Intel, IBM, and MS should sat on their hands and waited until the 386 was available. Of course, they wouldn't be any 386 under that scenario. We'd probably be still building computers kits or buying Apple IIs.
It's easy to pass judgement from the perspective of today's bloated PCs. Call me when Linux is ported to the Apple II.
This may be the first time on Slashdot that a comment about how people are going adopt a product that a vendor has explicitly stated they're not going to make has been modded as "Insightful".
Apple is not going to offer OS X as a seperate product, their not going to offer any dual-boot configurations and they'll sue anyone who tries to sell one. It's as simple as that.
"Well, I'll buy the Apple because they make good machines, and if OS X is crap then I'll just install Windows."
So customers that never had a Mac know that Macs are good machines but didn't want to buy one until they could install Windows on it just in case they're wrong. Talk about a convoluted argument!
Even if someone illegaly modifies OS X to run on a PC it's unlikely that it will run on older machines. Since Apple doesn't have to worry about backward compatibility of hardware, they're going to create a computer with competitive specs, their not going to target 2004 class technology.
There's a big difference between using Windows and believing that it's better than Linux. If a majority of Slasdotters do believe that, they must be the ones who don't post or moderate (with a few exceptions).
Try this experiment: Make a mild comment about Windows being better than Linux. Then a few days later post the exact same comment with the words Linux and Windows swapped. See which comment is modded as Flamebait or Troll and which is modded Insightful or Informative.
Slashdot Mirror
"and if you were a developer chances are you'd rank Eclipse very high in your list of indispensable tools."
I guess you mean if I were an Open Source developer. I hate to break it to you, but this developer along with the majority of all developers don't use Eclipse at all.
But my point was that IBM makes money in so many ways that it's hardly a test case for the business viability of Open Source.
Red Hat is a better example but probably still hasn't made a profit overall. That's pretty amazing considering that 90% of their development was done for free.
I'll be convinced when I see companies make a profit creating orignal software that is completely open and without any proprietary extensions.
"Counter-example: IBM, a heavy-weight in open-source, is not starving."
Yes, with their hardware sales and proprietary software sales, they're doing quite well. What was your point again?
I think the writers ran out of original ideas for parallel worlds so it ended up being the Mag show.
"Yes a bunch of guys running Autocad might have an issue and it probably wouldn't make sense to switch them. But are they really the users still running Win2K? "
Actually, I have archietect friend who's company was running Windows 95 and the latest version of AutoCad a year ago. They've upgraded to Windows XP now, but the point is that if they stayed with Windows 95 so long, it's plausible that other firms using AutoCad might be sticking with Win2k.
"But to suggest none of the products previously list will be Open Sourced is one heck of a prediction."
IBM refuses to open source some products like Visual Test that they no longer sell, so it's one heck of a prediction that any big money maker will be open sourced as long as it's still making money.
"Why, because if you don't you are missing out on developer mind share."
You're assuming that all developers care about a product being F/OSS. At least today, the vast majority of development is being done on proprietary products.
"I avoid proprietary software for moral reasons. I expect corporate support for my ideology based on sensible business practices. "
So if MS avoids F/OSS for sensible business practices you shouldn't complain I guess.
"You can bet however that Big Blue will Open source thier big money spinners as soon as it does make sense (read will make them more money)."
You can also bet that the day it makes sense to IBM will never come. I'll bet that IBM has plans for proprietary products that they haven't even started working on yet.
In fact, here is a great question to ask IBM executives. "Will you commit to making all new software products open source by the end of this decade?"
I'll bet the answer will not be a simple "Yes".
I think you're jumping to a conlusion here.
First, you're assuming that those customers don't like Windows XP as opposed to simply chosing to stay with a version that currently meets their needs.
Second, you're assuming that if they finally feel forced to upgrade because support is ending they'll choose an alternative that isn't supported at all, doesn't run their current applications and requires a steep learning curve to use.
If they like Windows 2000 then they're certainly going to like Windows XP more than Linux.
The spin on this story seems to be that MS will be hurt because their customers aren't upgrading to Win XP.
The other fact that this story reveals is that many MS customers are so happy with Win 2K that they don't want to change. That inertia is far more damaging to the prospect of Linux on the desktop than it is to MS's bottom line.
They should also open Visual Test since they will no longer allow you to buy a license for it.
"WebSphere Application Server is the market leading application server, I'm pretty sure. IBM eclipsed (no pun intended) BEA some time ago."
I guess the Application Server market is defined as Internet applications running on non-Apache web servers.
Overall, it still looks like IBM isn't an undisputed market leader in any category and a lot of their success seems to revolve around acquisitions rather than internal development.
I'm well aware of IBM's purchase of Rational since it resulted in the abandonment of Rational Visual Test customers (including myself and my clients). But buying software companies is not an indication that you know how to develop software. Quite the opposite I would argue.
"As I pointed out very early on in this, er, discussion, BIOS calls would be problematic, but even so, a file system with decent (and reserved) permissions would have gone a long way to discouraging a lot of bad things."
There were three common ways to program the PC, use the OS, use the BIOS and do your own thing. Very few successful commerical applications used the OS method exclusively because it was the slowest way.
If your intent all along was to say that DOS was unstable and insecure with respect to how it enforced the rules for applications that only used the OS, then I partially agree with you. Assuming they had sufficient resources they could have made the file system more secure as you suggest.
Note however that the key stability issue still couldn't be solved by the OS because application bugs could still have pointers pointing to any memory location and the 8088 didn't have any mechanism for restricting access.
"The fact that DOS on a 386 was just as promiscuous as DOS on a 8088 just proves my point that security was not a design consideration for MS, even when the hardware tools were available and waiting to be used."
This is a bit of goalpost moving on your part. The issue was whether the original design of DOS was less secure than it could have been given the limitations of the orginal PC. The 386 wasn't even in the design phase at that time.
"However, you and the other guy using your UID are having your differences, so I'll leave you two to finish the debate."
Cute but not exactly a graceful response to an apology.
What software product does IBM make that is a market leader? Most of their software isn't even ranked #2.
So it makes sense for IBM try other approaches to improve their market position. It would be a mistake for the market leaders to drop their more effective processes and adopt IBM's revised techniques until IBM can prove their value in the market.
I didn't read your post carefully enough before I made my last comment. Perhaps you do understand how to do things without involving the OS, so my comment about your skills was out of line. It's just your conclusions that are I believe are wrong.
It's true that DOS on the 386 was probably no more secure than DOS on the 8088 because MS didn't rewrite DOS to take advantage of the 386 capabilities. But that doesn't prove that one can write a secure and stable OS for the 8088.
On 8088-based machine the only way you could insure that self-modifying code was not possible would be to have the code in read-only memory. On the other hand, on a 386 the OS could setup the memory management in such a way that an application that is loaded doesn't have write access to its own code space.
Likewise, as long as interrupt vectors are stored in RAM they can be hijacked on the 8088 regardless of the OS.
I suspect (based on your description of your experience) that you already know much of what I explained here but don't want to admit that the original DOS wasn't that bad given the limitations of the orginal PC and that the primary techniques used to get around DOS could have been used against any OS in an 8088-based machine.
"And which was more stable and capable than (quick and dirty) MS-DOS, backing up my claim."
So your claim is that CP/M-86 which could have been implemented for the PC but wasn't, would have been more stable and capable than MS-DOS which was. Sounds speculative to me. So explain to me exactly how CP/M-86 would have been more stable than MS-DOS given the fact that the 8088 would still allow any application to access any memory and write to any hardware device without permission of the OS.
It wasn't my intention to misrepresent your argument or put words in your mouth but you do keep trying to bring this argument forward to the present with things like: "Which has been the design principle that MS has been stuck with ever since." I was merely anticipating an argument you I thought you were hinting at. If that was unfair, I'm sorry.
"Huh? You could overwrite system files, write self-modifying code, build code on the fly in the data segment and run it from there, steal interrupt vectors, modify the MBR, the list is nearly endless."
First of all, you were talking about users, not programmers.
If you really wrote code for the orginal PC and think you couldn't do those things without cooperation from an OS, you weren't very good at it.
As I asked before, please explain what technique a 8088-based OS could do to prevent (for example) an application program from overwriting a system file when the hardware allows writes to the disk directly without the OS being involved.
"My alternative: turn on, type URL of site. (GASP! Sounds like a "hard to use" CLI!!)"
If you're going to turn this into a GUI vs. CLI argument then by all means use the console browser. You want to go back to the previous page, type "b a c k". Gee I wonder what a web site would look like with pictures?
"The MS policy, until very recently, has always been to let a user modify anything."
There was very little that a user could modify in DOS. But my main point was that basic security and stability must be supported by the hardware; which the OS can take advantage of or not. The 8088 doesn't have such features, so applications can ignore the OS altogether and access hardware any way they please. This would be just as true for Unix as it was for DOS.
"Although BIOS calls would be a problem, any group of CS students could write an 808x OS more secure than MS-DOS."
If by this you mean that they could write a more secure OS for the original PC configuration (which we've been talking about), I doubt it. Most CS students would find it a major challenge to write any OS for that platform and for good reason.
You'd have to write most if not all of it in assembly. I suspect that most CS students spend very little time writing assembly language programs because so much of industry work is done in high level languages. That's not to say that they're not smart enough, they just don't have the training for it.
"Again, don't lecture me on MS security until they allow all users to do security updates on their systems."
I'm not lecturing you on MS security in general, but I am pointing out some misconceptions that you have about early PCs.
OK, so CP/M-86 could have been the PC's OS. I never claimed that DOS was the only possible PC OS IBM could have used, just that the limitations of the orginal PC precluded using an OS that we would consider stable and secure by today's standards.
"it will run anything for any reason at any time"
Please explain to me how you implement privilege levels in an OS on a processor that doesn't support it. All programs running on an 8088 have full access to the entire address space and there's nothing an OS can do about it.
What reliable secure contemporary of DOS and early Windows ran on a PC? In any case, I wasn't providing an excuse for Windows based on its roots, I was responding to the claim that MS had historically always done a poor job. The fact was that more capable OS's were MIA at the time simply because they couldn't be implemented on the platform or wouldn't be any more stable if they had.
"That's been the problem with MS operating systems and the MS philosophy all along."
Well, you can't implement a secure and stable OS like Linux on a 8088 with 16K or RAM, so Intel, IBM, and MS should sat on their hands and waited until the 386 was available. Of course, they wouldn't be any 386 under that scenario. We'd probably be still building computers kits or buying Apple IIs.
It's easy to pass judgement from the perspective of today's bloated PCs. Call me when Linux is ported to the Apple II.
This may be the first time on Slashdot that a comment about how people are going adopt a product that a vendor has explicitly stated they're not going to make has been modded as "Insightful".
Apple is not going to offer OS X as a seperate product, their not going to offer any dual-boot configurations and they'll sue anyone who tries to sell one. It's as simple as that.
"Well, I'll buy the Apple because they make good machines, and if OS X is crap then I'll just install Windows."
So customers that never had a Mac know that Macs are good machines but didn't want to buy one until they could install Windows on it just in case they're wrong. Talk about a convoluted argument!
Even if someone illegaly modifies OS X to run on a PC it's unlikely that it will run on older machines. Since Apple doesn't have to worry about backward compatibility of hardware, they're going to create a computer with competitive specs, their not going to target 2004 class technology.
There's a big difference between using Windows and believing that it's better than Linux. If a majority of Slasdotters do believe that, they must be the ones who don't post or moderate (with a few exceptions).
Try this experiment: Make a mild comment about Windows being better than Linux. Then a few days later post the exact same comment with the words Linux and Windows swapped. See which comment is modded as Flamebait or Troll and which is modded Insightful or Informative.