If the point of this document is to show that the Linux community is just as capable of generating FUD as Microsoft, then it has succeeded. As a tool for realistically positioning Linux's capabilities, though, it is quite useless.
This is some of the most half-baked blathering I have seen in a very long while, and it is really quite sad: every fool who goes in making claims like these and is promptly shot out of the saddle will set the movement back in the eyes of those who watch it happen. If the Linux community is going to get serious about taking on Windows NT, it will have to do a lot better than this.
How many "common customers" use 4-way NT boxes? Very few, in my experience.
This statement is a brilliant testimonial to the sheer naivite of Linux advocates, showing a very deep misunderstanding of how servers are deployed and used. This kind of comment strongly supports the view that many Linux developers are holed up in a bedroom somewhere, tweaking code on a home PC in their spare time.
NT loses these same benchmarks when comparing single-to-single processor and dual-to-dual processor machines.
And which benchmarks are these? I didn't see a footnote here.
Microsoft also claims that NT performs better than Linux when serving static web pages. However, e-business is not powered by static pages. It's powered by Active Server Pages and CGI5. Windows doesn't fare nearly so well in these comparisons.
Again, where is the evidence for this claim?
As for the technical specs quoted by Microsoft... they are out of date. The Linux kernel addresses 4 GB RAM, not two.
The 4 GB function is a kernel patch to the 2.3 kernel, which means that it is non-production, beta code until 2.4 ships. If I buy Red Hat 6.1 or some other 2.2-based kernel today, it will support no more than 2 GB...period.
Among the file systems Linux supports is SGI's XFS, recently released to Open Source, with a max file size of nearly one million terabytes
Again, this is not production code...it is a statement of intention by SGI to release its XFS code to the Linux development community.
Note that not all of the features supported by Linux are included by default in every distribution, but they all can be added if missing.
OK, and who will support these functions when they are added? The distribution suppliers? Not if it isn't in their product. The hordes of volunteer help-desk personnel idling on USENET groups? Only if it's a K00l question, dude. The profusion of promising startups dedicated to supporting commercial Linux sites? Better check that fine print again...!
The fact that this can be compiled into the OS kernel or not, depending on the needs of the users allows every installation to tailor the smallest, fastest, most stable custom kernels to their specific needs.
Assuming they know how to compile a kernel...
Linux's stability is only based on anecdotes. Microsoft seems unable to differentiate anecdotes from testimonials.
OK, and where are these "testimonials" again (and I'm not counting USENET or Slashdot postings)?
Be that as it may, there are a lot of these anecdotes. Many of them are include documented uptimes ranging from months to years. Footnote: I myself experienced an eight-month uptime between kernel upgrades, and I do not mean scheduled uptime.
See above.
Microsoft also claims that Linux has no journaling file system, ignoring the fact that the SGI's XFS is a journaling file system
Again, this is not production code.
They also ignore the fact that NT 4.0 itself lacks a journaling file system!!
This is just wrong, and you are not doing the Linux community any favors by essentially lying about Windows NT's capabilities. Here's a simple test for you: put two identical systems next to each other, one running Windows NT 4.0 and the other Linux. Boot each of them up. Then pull both power plugs out, and reinsert them at the same time. Which system will be up and running faster?
Also, Windows NT clustering is limited to failover ONLY.
This is also wrong. Windows NT 4.0 has Web-server load balancing functions built in as part of its Convoy clustering technology. Oracle Parallel Server is widely deployed on Windows NT to create scalable database clusters.
Linux clustering was developed in association with NASA, an agency having a far stricter definition of "mission criticality" than any commercial entity. NT has no equivalent technology.
This is wrong again. The PVM and MPI technologies on which Beowulf is based have been running -- and have been widely deployed -- on NT for years.
Free doesn't mean low TCO. Actually, it does. Microsoft's TCO calculations are based against other commercially marketed Unixes, which have very expensive initial acquisition and support contract costs and traditionally high education costs.
And Linux support contracts are free? There won't be any education costs to move users to Linux?
Support costs can be very low, as purchased support can be supplemented with award-winning Usenet support.
I'm sorry, I am picking myself up off the floor from laughing so hard. I have submitted many, many questions to various Linux discussion groups over the years and the quality of answers is *wildly* uneven, with the majority of answer ranging from irrelevant to plain old wrong. Any IT professional who depends solely on USENET for Linux support should be fired.
Microsoft itself charges for support for a product that they licensed (not sold!) you at considerable cost without warranty.
And Linux support programs *do* come with a warranty? Please, show me these programs!
However, with Linux you're not required to purchase support at all.
With Windows, you're not required to purchase support either. What's your point?
The Linux User community, operating at no charge, garnered the 1996 InfoWorld Product of the Year award for Best Technical Support.
InfoWorld has a long history of anti-MS sloganeering, and seems to give out its awards simply based on the fact that they aren't MS products. After all, this is the magazine that for the previous five years had given OS/2 their product-of-the-year award. If I as an IT professional had made a purchasing decision based on InfoWorld's recommendations in 1994, I would be in fairly deep trouble right now.
Microsoft claims that your security administrator must be an expert to properly configure security. My own knee-jerk reaction to this is, "When do you NOT want an expert supporting your systems? If you can't afford one full-time you hire a consultant"
OK, and you just got through going on about how low Linux cost-of-ownership was?
In point of fact, NT's security is... not too difficult to crack.
OK, this is is just plain FUD (meaning you have made a claim that is supported by nothing more than the fact that you made the claim).
Properly configuring Linux security is mostly a matter of removing those services that are not needed and religiously applying security patches as they appear. This should be standard operating procedure for any, regardless of the platforms used.
Correct, so why should it be any different for NT?
Microsoft claims that existing Linux GUIs are cumbersome and difficult to use. In fact, my mother sat down and began using the KDE desktop with no training, no prior experience, and not one single problem.
OK, now let's put her in front of a GNOME desktop with a stopwatch, and see how long it takes her how to figure out how to do things there.
For business use, the major general purpose tool Linux lacks at the moment is a Lotus Notes client.
And an Exchange client (although that is hardly surprising). So now we have taken about 90% of the messaging users off the table.
And the Windows Notes client can be run if you simply install WINE (Windows emulator for Linux).
And just how long will the program run before a segmentation fault?
The Future.....
In the future, we'll all be flying around in air cars! Nuff' said.
I was a little disappointed to see Alan jump into this debate. I believe his strengths as a technician are unassailable, and that his greatest value to the community is to make Linux the best technical solution, period - which simply makes all the other questions about open vs. close source, NT vs. UNIX, Microsoft etc. moot.
Aside from the fact that the computer-car analogy is one of the most overused analogies in existence, he does not add much value to the open source side of the debate.
There are a few exceptions. For example, the comment about NT-Alpha users is dead-on! I'll bet *their* opinion on open source is changing dramatically right now.
But he has trivialized the other questions, which are primarily business issues that are far more subtle and complex than he presents. For example, he questions why anyone (i.e. Windows users) would commit to a single source for support and parts. Well, the fact is, businesses do *not* usually make their commitment to Microsoft, they make it to Microsoft's partners, for which there is a profusion of choice! If the Windows NT support from Dell or my reseller is insufficient, I can go to Unisys. If that doesn't work out, I can go to IBM or Compaq. The same goes for buying a Windows system. Granted, Microsoft lies *behind* all of these entry points and is ultimately responsible for resolving issues, but that fact rarely becomes visible to the user. From the user's standpoint, there are enough choices that he can worry about what matters - running his business. Now, of course, Linux provides far more flexibility for the *suppliers*, which is why the various OEM agreements being put into place are the most significant evidence of Linux's progress. However, that has little to do with decisions made by users - most will accept what the system vendors provide, as long as there is choice at the levels that matter: hardware and support.
Alan should stick with what he does best: making Linux real. Everything else just adds to the noise.
As this article correctly points out, the failure occurred in the application -- not Windows NT itself. The divide-by-zero condition caused the program to go into a tight loop, which made it unresponsive to other applications across the network that depended on it. The result was a chain reaction that evidently froze control of the propulsion system. Also, if I remember correctly, the failure of the NT application did *not* cause the ship to be towed into port, although it did force the ship to idle at sea for several hours while the network was rebooted -- the towing incident resulted from some other, unspecified failure related to the use of on-board LAN technology, in which NT may or may not have been the culprit.
Shortly after these incidents, they were reported in a single article (Computer Reseller News, I believe) at moderately competent detail, based largely on the complaints of the "whistle blower" mentioned in this article as well. But that one CRN article was summarized with diminishing fidelity in several other editorials, which were then widely replicated thanks to the magic of syndication, and the incident eventually became the part of Windows folklore in mutated form, along the lines of "the main server blue-screened when the navigator punched in a new course".
It hardly seems fair to blame NT for an application failure. Moreover, the complaints about the Navy cutting costs by using NT seem targeted instead at the broader isssue of replacing expensive, but reliable, fault-tolerant systems with commodity and LAN-based technology, rather than an attack on NT itself. As described, this type of failure could have happened just as easily if the on-board systems had been running Linux.
Increased software patent registration and enforcement are simply the cost of Open Source growth. In the past, with closed source, software patents were far less of an issue, because there was usually no way to tell how a program did something. Now that we agree the internals of software should be exposed for all to examine and improve upon, there will have to be formal ways for designers to benefit from their innovations. You can't have it both ways, i.e. open access to all source code *and* no patent protection, unless you believe that the entire concept of intellectual property is obsolete, and I think that is a stretch by all but the most extremist views.
Algorithms are not "mathematical formulas"...they are *structures* that are composed from primitives (i.e. the basic instructions of various programming languages), and are thus no different than the mechanical assemblies patentable under traditional laws. To use the popular metaphor of Open Source allowing you to "open the hood of the car" that would be "welded shut" with closed source, you can now see the engine, but that doesn't mean you have the right to build and sell an exact copy yourself. You *do*, however, have the right to look at the engine and improve its design. In the case of the example given in the paper, if you do not have the right to use patented "Save As..." methods, well, go ahead and invent a *better* way to initiate a write to disk. Thus, it seems to me that software patents could encourage innovation, rather than inhibit it.
Having said this, I do agree that the problem of *correctly* enforcing software patents is non-trivial, and will require significant effort to address. Since it all comes down to "prior art", the PTO must improve its processes so that authorities can recognize such instances efficiently.
Slashdot Mirror
If it isn't there, well... then MS will lose the server market completely in no time.
Well, Linux doesn't have journaling now, and it hasn't "completely lost the server market", so why would that be true for NT?
If the point of this document is to show that the Linux community is just as capable of generating FUD as Microsoft, then it has succeeded. As a tool for realistically positioning Linux's capabilities, though, it is quite useless.
... not too difficult to crack.
This is some of the most half-baked blathering I have seen in a very long while, and it is really quite sad: every fool who goes in making claims like these and is promptly shot out of the saddle will set the movement back in the eyes of those who watch it happen. If the Linux community is going to get serious about taking on Windows NT, it will have to do a lot better than this.
How many "common customers" use 4-way NT boxes? Very few, in my experience.
This statement is a brilliant testimonial to the sheer naivite of Linux advocates, showing a very deep misunderstanding of how servers are deployed and used. This kind of comment strongly supports the view that many Linux developers are holed up in a bedroom somewhere, tweaking code on a home PC in their spare time.
NT loses these same benchmarks when comparing single-to-single processor and dual-to-dual processor machines.
And which benchmarks are these? I didn't see a footnote here.
Microsoft also claims that NT performs better than Linux when serving static web pages. However, e-business is not powered by static pages. It's powered by Active Server Pages and CGI5. Windows doesn't fare nearly so well in these comparisons.
Again, where is the evidence for this claim?
As for the technical specs quoted by Microsoft... they are out of date. The Linux kernel addresses 4 GB RAM, not two.
The 4 GB function is a kernel patch to the 2.3 kernel, which means that it is non-production, beta code until 2.4 ships. If I buy Red Hat 6.1 or some other 2.2-based kernel today, it will support no more than 2 GB...period.
Among the file systems Linux supports is SGI's XFS, recently released to Open Source, with a max file size of nearly one million terabytes
Again, this is not production code...it is a statement of intention by SGI to release its XFS code to the Linux development community.
Note that not all of the features supported by Linux are included by default in every distribution, but they all can be added if missing.
OK, and who will support these functions when they are added? The distribution suppliers? Not if it isn't in their product. The hordes of volunteer help-desk personnel idling on USENET groups? Only if it's a K00l question, dude. The profusion of promising startups dedicated to supporting commercial Linux sites? Better check that fine print again...!
The fact that this can be compiled into the OS kernel or not, depending on the needs of the users allows every installation to tailor the smallest, fastest, most stable custom kernels to their specific needs.
Assuming they know how to compile a kernel...
Linux's stability is only based on anecdotes. Microsoft seems unable to differentiate anecdotes from testimonials.
OK, and where are these "testimonials" again (and I'm not counting USENET or Slashdot postings)?
Be that as it may, there are a lot of these anecdotes. Many of them are include documented uptimes ranging from months to years. Footnote: I myself experienced an eight-month uptime between kernel upgrades, and I do not mean scheduled uptime.
See above.
Microsoft also claims that Linux has no journaling file system, ignoring the fact that the SGI's XFS is a journaling file system
Again, this is not production code.
They also ignore the fact that NT 4.0 itself lacks a journaling file system!!
This is just wrong, and you are not doing the Linux community any favors by essentially lying about Windows NT's capabilities. Here's a simple test for you: put two identical systems next to each other, one running Windows NT 4.0 and the other Linux. Boot each of them up. Then pull both power plugs out, and reinsert them at the same time. Which system will be up and running faster?
Also, Windows NT clustering is limited to failover ONLY.
This is also wrong. Windows NT 4.0 has Web-server load balancing functions built in as part of its Convoy clustering technology. Oracle Parallel Server is widely deployed on Windows NT to create scalable database clusters.
Linux clustering was developed in association with NASA, an agency having a far stricter definition of "mission criticality" than any commercial entity. NT has no equivalent technology.
This is wrong again. The PVM and MPI technologies on which Beowulf is based have been running -- and have been widely deployed -- on NT for years.
Free doesn't mean low TCO. Actually, it does. Microsoft's TCO calculations are based against other commercially marketed Unixes, which have very expensive initial acquisition and support contract costs and traditionally high education costs.
And Linux support contracts are free? There won't be any education costs to move users to Linux?
Support costs can be very low, as purchased support can be supplemented with award-winning Usenet support.
I'm sorry, I am picking myself up off the floor from laughing so hard. I have submitted many, many questions to various Linux discussion groups over the years and the quality of answers is *wildly* uneven, with the majority of answer ranging from irrelevant to plain old wrong. Any IT professional who depends solely on USENET for Linux support should be fired.
Microsoft itself charges for support for a product that they licensed (not sold!) you at considerable cost without warranty.
And Linux support programs *do* come with a warranty? Please, show me these programs!
However, with Linux you're not required to purchase support at all.
With Windows, you're not required to purchase support either. What's your point?
The Linux User community, operating at no charge, garnered the 1996 InfoWorld Product of the Year award for Best Technical Support.
InfoWorld has a long history of anti-MS sloganeering, and seems to give out its awards simply based on the fact that they aren't MS products. After all, this is the magazine that for the previous five years had given OS/2 their product-of-the-year award. If I as an IT professional had made a purchasing decision based on InfoWorld's recommendations in 1994, I would be in fairly deep trouble right now.
Microsoft claims that your security administrator must be an expert to properly configure security. My own knee-jerk reaction to this is, "When do you NOT want an expert supporting your systems? If you can't afford one full-time you hire a consultant"
OK, and you just got through going on about how low Linux cost-of-ownership was?
In point of fact, NT's security is
OK, this is is just plain FUD (meaning you have made a claim that is supported by nothing more than the fact that you made the claim).
Properly configuring Linux security is mostly a matter of removing those services that are not needed and religiously applying security patches as they appear. This should be standard operating procedure for any, regardless of the platforms used.
Correct, so why should it be any different for NT?
Microsoft claims that existing Linux GUIs are cumbersome and difficult to use. In fact, my mother sat down and began using the KDE desktop with no training, no prior experience, and not one single problem.
OK, now let's put her in front of a GNOME desktop with a stopwatch, and see how long it takes her how to figure out how to do things there.
For business use, the major general purpose tool Linux lacks at the moment is a Lotus Notes client.
And an Exchange client (although that is hardly surprising). So now we have taken about 90% of the messaging users off the table.
And the Windows Notes client can be run if you simply install WINE (Windows emulator for Linux).
And just how long will the program run before a segmentation fault?
The Future.....
In the future, we'll all be flying around in air cars! Nuff' said.
I was a little disappointed to see Alan jump into this debate. I believe his strengths as a technician are unassailable, and that his greatest value to the community is to make Linux the best technical solution, period - which simply makes all the other questions about open vs. close source, NT vs. UNIX, Microsoft etc. moot.
Aside from the fact that the computer-car analogy is one of the most overused analogies in existence, he does not add much value to the open source side of the debate.
There are a few exceptions. For example, the comment about NT-Alpha users is dead-on! I'll bet *their* opinion on open source is changing dramatically right now.
But he has trivialized the other questions, which are primarily business issues that are far more subtle and complex than he presents. For example, he questions why anyone (i.e. Windows users) would commit to a single source for support and parts. Well, the fact is, businesses do *not* usually make their commitment to Microsoft, they make it to Microsoft's partners, for which there is a profusion of choice! If the Windows NT support from Dell or my reseller is insufficient, I can go to Unisys. If that doesn't work out, I can go to IBM or Compaq. The same goes for buying a Windows system. Granted, Microsoft lies *behind* all of these entry points and is ultimately responsible for resolving issues, but that fact rarely becomes visible to the user. From the user's standpoint, there are enough choices that he can worry about what matters - running his business. Now, of course, Linux provides far more flexibility for the *suppliers*, which is why the various OEM agreements being put into place are the most significant evidence of Linux's progress. However, that has little to do with decisions made by users - most will accept what the system vendors provide, as long as there is choice at the levels that matter: hardware and support.
Alan should stick with what he does best: making Linux real. Everything else just adds to the noise.
As this article correctly points out, the failure occurred in the application -- not Windows NT itself. The divide-by-zero condition caused the program to go into a tight loop, which made it unresponsive to other applications across the network that depended on it. The result was a chain reaction that evidently froze control of the propulsion system. Also, if I remember correctly, the failure of the NT application did *not* cause the ship to be towed into port, although it did force the ship to idle at sea for several hours while the network was rebooted -- the towing incident resulted from some other, unspecified failure related to the use of on-board LAN technology, in which NT may or may not have been the culprit.
Shortly after these incidents, they were reported in a single article (Computer Reseller News, I believe) at moderately competent detail, based largely on the complaints of the "whistle blower" mentioned in this article as well. But that one CRN article was summarized with diminishing fidelity in several other editorials, which were then widely replicated thanks to the magic of syndication, and the incident eventually became the part of Windows folklore in mutated form, along the lines of "the main server blue-screened when the navigator punched in a new course".
It hardly seems fair to blame NT for an application failure. Moreover, the complaints about the Navy cutting costs by using NT seem targeted instead at the broader isssue of replacing expensive, but reliable, fault-tolerant systems with commodity and LAN-based technology, rather than an attack on NT itself. As described, this type of failure could have happened just as easily if the on-board systems had been running Linux.
Increased software patent registration and enforcement are simply the cost of Open Source growth. In the past, with closed source, software patents were far less of an issue, because there was usually no way to tell how a program did something. Now that we agree the internals of software should be exposed for all to examine and improve upon, there will have to be formal ways for designers to benefit from their innovations. You can't have it both ways, i.e. open access to all source code *and* no patent protection, unless you believe that the entire concept of intellectual property is obsolete, and I think that is a stretch by all but the most extremist views.
Algorithms are not "mathematical formulas"...they are *structures* that are composed from primitives (i.e. the basic instructions of various programming languages), and are thus no different than the mechanical assemblies patentable under traditional laws. To use the popular metaphor of Open Source allowing you to "open the hood of the car" that would be "welded shut" with closed source, you can now see the engine, but that doesn't mean you have the right to build and sell an exact copy yourself. You *do*, however, have the right to look at the engine and improve its design. In the case of the example given in the paper, if you do not have the right to use patented "Save As..." methods, well, go ahead and invent a *better* way to initiate a write to disk. Thus, it seems to me that software patents could encourage innovation, rather than inhibit it.
Having said this, I do agree that the problem of *correctly* enforcing software patents is non-trivial, and will require significant effort to address. Since it all comes down to "prior art", the PTO must improve its processes so that authorities can recognize such instances efficiently.