> RMS can tell you it already has done so ["crush" OSS], back in the late 70's/early 80's. Not by force, but by hiring away the best and brightest.
Well, it clearly didn't do it very thoroughly.... In fact, what it [business] did was break up the AI Lab at MIT in the midst of the "artificial intelligence" marketing craze of that era, the sole long-term effect of which was to offend RMS' communal sensibilities (understandably enough) and inspire him to dedicate his considerable hacking talents to e.g. gcc and other utilities central to OSS. Prolog and its friends are at best confined to a tiny niche; OSS is in the process of squashing NT, Microsoft's Prime Weapon for dominating enterprise computing. Some "crush."
But in spite of RMS' offended spirit, open software continued to be available from numerous other sources, both in academia (think about BSD and its associated stuff) and elsewhere. The "growth of the internet" by any reasonable metric started around 1983; basic free stuff -- like sendmail, vi, and so on -- were circulating from the very beginning. In 1991, when Linus first put up version 0.00whatever, the internet was a tiny fraction of its current size -- but it was big enough....
RMS himself, by the way, points out that nearly two-thirds of a standard Linux distribution consists of non-GPLed software, so apparently the FSF, for all its valuable contribution to OSS, is still in the minority when it comes to defining the term "free software."
> They are not drones whose vision of life exclusively consists of "graduate and get a [gasp!] job!".
I don't know anyone whose vision of life consists exclusively of anything. Do you?
Likewise I know very few people whose vision of life does not include food and shelter, though in some cases (notably RMS') in a rather minimalistic form.
> You assume that "enlarging" is a good thing. Why?
Gee, I dunno. I just looked around and noticed that a hell of a lot more great software with source was available now than it was when I first started using it nearly ten years ago. Silly me, I thought it might be because more people were writing the stuff....
> You also assume that some sort of plusralism in the motivations of a group is all right. Why?
I assume it's all right for the same reason I assume gravity is all right; it's inescapable. "A group" is not a thing, it's a word for a number of people who choose to cooperate with each other in some area of endeavor. Show me somewhere a group all of whose members are there with the same motivations, values, and goals and I'll show you a group that is at the very least unlikely to be individualistic enough to do any really superb hacking....
> Is it the freedom to change Linux into a system that relies upon utilities and applications that are not as Free as those distributed under the GPL?
The last time I looked, Linux depended for its GUI, for example, on X. Is that "more" free, "less" free, or "as" free as if X(Free!) were under the GPL?
Note that the definition of "free" promulgated by the GPL is oddly one-sided; many of us object to the habitual use of "free software" as a synonym for "GPLed software", just as we'd object to restricting the phrase "Let's do lunch" (now trademarked by Frito-Lay) to proposal of occasions when we planned on consuming potato chips.
> Every new technology is marked by dreams of freedom. Seldom are these dreams fulfilled.
Hmmm. The near-ubiquity of transistor radios in the Third World, the green revolution that showed up all the Malthusian doomsday scenarios of the '70s for the nonsense they were, the Ford Model T that set in motion the incredible mobility of the rural population, the ability of a quarter of a billion people or so all over the planet to communicate with each other instantaneously via the Internet, the revolution in "freedom of the press" brought on by the Web -- but these dreams are "seldom fulfilled." OK, sure.
If you're saying "Nothing is a panacea", that's true but trivially obvious. If you're saying any more than that, you're obviously quite wrong. Technological development in the last 150 years has vastly increased the empowerment of individuals.
... and in fact, if anything you understate the case.
Microsoft's habit of casually diddling the core of the operating system every time they need something new for marketing reasons has lead to fragmentation even within nominally identical platforms.
For example, if you apply Service Pack 1 to Visual C++ 5 and recompile your application, you discover that suddenly your code won't load on a machine running the original Win95 Retail. Why? Because SP1 changed a header file to default to using routines from a version of comctl32.dll -- one of Win32's most central libraries, the Common Controls library -- that was distributed only with Internet Explorer 4! You can fix your code once you've discovered the problem, of course, but documentation for this "upgraded feature" is totally buried in an obscure section of the documentation, and isn't even mentioned in the release notes!
Likewise many sites have found NT4 Service Pack 4 to break a number of crucial apps. So the "unity" of the Win32 platform is largely illusory anyway; at Microsoft, the perceived needs of the marketing department so far outweigh any technical considerations and user requirements that there's no contest.
And, when all is said and done, this is really what's hanging Microsoft. IT professionals all over the planet are sick of these marketing games and the resulting shoddy code; they're eager -- almost desperate -- for an alternative to Microsoft in the server room (which is where their most critical problems originate). And this is why they're leaping into Linux (and FreeBSD and even NetWare (again)) with glad cries and tears of joy and relief.
> But can you run Solaris/x86 binaries on a SCO box?
No, but why would you want to? The whole "shrink-wrap" ethos is basically an outgrowth of the PC explosion.
Note that game suppliers and the packagers of bundled software for (e.g.) scsi scanners simply provide both Win and Mac versions on the distribution CD -- and at this point differences between SCO, Linux, Solaris, AIX, and HP-UX sources are so well-known that operating from a common source base isn't at all difficult -- and the "intellectual property" concern leading to the distribution of binary-only code is based on the source, anyway.
One of the design mistakes made with OS/2 -- or at least one of the things that hampered its spread -- was that you couldn't simply recompile a Win application for it. IBM tried to remedy that, to a limited extent, with Merlin's compatibility APIs, but it was too late. Although you never know, it may yet recover in the current wave of interest in Microsoft alternatives; reports of its death have been greatly exaggerated before....
Someone posting illiterately as an AC who actually expects to be taken seriously when he accuses others of being "wannabes". How did you get in here, sonny? Didn't Rob card you? Or did you tell him you were looking for your daddy?
> Can't Linux just balance the processes between two processors?
Yeah, and he does it pretty well. But keeping both processors busy on a desktop machine is a good trick; most of us are doing only one really heavy computing task at a time; downloading in the background hardly keeps the CPU awake, let alone busy.
But if you make -j3 your kernel on an SMP machine, for example, which tells make to spawn and manage three separate jobs, you'll notice quite a speedup; you'll save about a third of the time (not half, because of overhead) if you've got enough memory.
Well, since Mozilla apparently involves a complete rewrite of the rendering engine and conversion of the UI to GTK+, it should be a) a lot snappier and b) less trouble-prone -- unless the actual source of our problems is its net handling, which I doubt.
By the way, kfm's html handling is much improved in KDE 1.1; try browsing with it now. Some stuff is still not quite right, but it's mostly very usable.
> ... SLOW.... it takes for ever for windows to repaint....
This is odd; my impression of GNOME was that it was if anything slightly snappier than KDE. (Both built on my RH 5.1+ machine, egcs -O6 -mpentium.) How much memory do you have? How much else is going on in the background?
Also check your XFree version -- 3.3.3.1 is a little better with the Matrox card than 3.3.2.
> it sez libjpeg.so.6 not found, but i have the 5.2's jpeg lib 6.2 installed. (libjpeg.so.62.)
Libjpeg is weird in that its header file contains both major and minor number together (61, 62, etc.) and it refuses to load without an exact match. The KDE devel crew uses.61 (I had this problem with GNOME, too, but they use.62 [sigh...].)
Try linking libjpeg.so.61 -> libjpeg.so.62, ldconfig, and see if it runs....
Alternatively: Get the kdesupport rpm, make sure libjpeg.so.61 installs in/opt/kde/lib, make sure/opt/kde/lib is mentioned in ld.so.conf, ldconfig, and see if it runs....
" Server busy or unable to fulfill request. The server is unable to fulfill your request due to extremely high traffic or an unexpected internal error. Please attempt your request again (if you are repeatedly unsuccessful you should notify the site administrator). (Location Code: 25)"
... is that IIS gets slashdotted a lot easier than its alternatives.... But we knew that...
I was under the impression that although C calling conventions were tightly specified in their type conversions, argument order, and stack handling, the layout of the C++ object's internal structure -- in particular, its method vector table -- was left up to the compiler implementor.
If I'm right about this, it would seem to me that adding language bindings to a C++ library would be substantially trickier than to a C library.
Myself, I prefer to work in C++, but de gustibus non est disputandum (except of course that COBOL sucks and RPG is a joke...).
On the other hand, I think I'm still being haunted by by the Ghost of GNOME Past, since I downloaded everything in source, followed the instructions on www.gnome.org with the care of a Buddhist monk pronouncing a Vedic hymn, and libgnomeui still won't build for me (an undefined symbol having to do with gtk_imlib or something; I'm at the wrong machine now. It's in a routine loading a png into a canvas, if I recall). It's probably due to some old header files or libs lying around in the wrong place, but I've looked for them and can't find 'em.
Over its history (and possibly still), GNOME stuff has lived in various places depending on when (0.2x vs now) and how (rpm vs tar.gz) it was installed. Does anybody have a script that will search and destroy all obsolete GNOME cruft everywhere?
Craig still using KDE... and xfce, and wmx, and WM...
Love says "everybody has to support the same version of the kernel and the same set of library services."
True but vacuous. Is there any general-purpose distro without plans to support 2.2.x, egcs, and glibc in the very near future?
The devil's in the details -- for example, the libjpeg problem with its unnecessarily fine-grained version code in its header. But these are easily solvable.
As for LSB, as long as it's a reasonable discussion between techies to solve technical problems, bless them. But -- at one point, at least -- it was in danger of becoming yet another vehicle for the overblown Perens ego, to which the various distributions replied (quite properly) by ignoring it and starting side discussions.
I have no idea why Love chose to write this now, unless there's some special situation at Caldera that requires a political statement of LSB support from them. I would have thought that the whole standards thing was mostly a non-issue by now....
machine with linux/Be $1000 hardware + $0 OS selling to consumer for $1200 OEM earns $200 profit/machine
... and if the oem doesn't sell the machine at all, he loses $1000.... and he's already lost his Microsoft bonus, so his Windows machines now cost him, say, $20 more apiece, which means he'll sell fewer of them. That's what the OEMs are afraid of.
Jon's windy self-important recycled Marxism takes me back to the days when scruffy passionate students would jump nimbly atop anything higher than a dictionary and deliver harangues of deeply-felt incoherent nonsense to other scruffy passionate students (including me, back before I traded in my hair for extra waistline).
They were complete imbeciles, of course. So is Jon. But a little nostalgia every now and then is good for the soul....
... you don't apparently understand why market solutions work best.
> 95% market share means it's the best product.
Nope. 95% market share means only that for 95% of the purchasers, it's an acceptable balance between functionality and cost on the basis of their individual knowledge. It says nothing whatever about inherent technical quality, aesthetic appeal, efficiency, usability, or anything else.
(And since when, by the way, do us libertarians, who profess a radical individualism, believe that the majority is necessarily right? Politically we reject the tyranny of the majority as firmly as we reject the Divine Right of Kings.)
Now, if you (or any of us Linux -- or Be, or OS/2, or whatever -- advocates) think they're wrong, the answer is to educate them, and do whatever is necessary to raise the functionality and/or lower the cost (not just $, but hassle and time) of your favorite OS. And I think that's what we're doing -- isn't it? -- with improved installation, KDE/GNOME, etc.? (Who have you introduced to Linux today?)
I'm not sure I understand why this is a problem, since Linux has included BSD networking code for ages and displays the "Regents of the University of California" message when it cranks up. Though perhaps making any program that uses the db module display the copyrights might be a drag...
The heedless fascism of the GPL has caused so much grief to the rest of the free software community, though, that I'm somewhat amused to see the FSF having similar problems with a much freer license....
> Who wants to pay attention to the nuts & bolts and the really CORE mechanics which is dull to say the least ??
Look at all the contributors to the kernel mailing list on such topics as memory management and file systems (are these "CORE mechanics" enough for you?).
Everything is fascinating to somebody; Alan Cox got hooked originally because he became fascinated by, of all things, ethernet cards.
And believe me, if there's a fundamental bug in the memory management or your ethernet card driver, whether these things interest you or not you'll notice it, and probably report it in a whiny and offended tone!
This silly piece -- insofar as it deserves any attention at all -- is subject to the same refutation used by Dr. Samuel Johnson against another silly theory in the late 18th century:
Of Bishop Berkeley's theory of the non-existence of matter, Boswell observed that though they were satisfied it was not true, they were unable to refute it. Samuel Johnson struck his foot against a large stone, till he rebounded from it, saying "I refute it thus."
That is, the writer asserts any number of things that might be true, but (empirically) aren't. Linux isn't showing up in server rooms all over the world because it's the Latest Hot Thing, one of the frequent fads that seem to sweep the IT management community; it's showing up because it is superior to all other available options for solving the problems the individual sysadmins must face.
If you actually analyze carefully Ted Lewis' article -- an exercise I don't recommend -- you'll see that none of his statistical arguments hold water -- even in terms of the mathematics, let alone their factual premises.
That is, where it isn't simply incoherent. Consider, for example, the following:
"As the
Linux 2.0: What It Is and Isn't sidebar shows, Linux has yet to incorporate many of the features and applications that will doom it to a complex future." [p. 127]
What does this sidebar actually say [p. 125]? That Linux has:
Multithreaded kernel
SMP
Multiple FS support
Disk striping and mirroring
Multiprotocol networking
X11 GUI
GNU tools
all the internet tools
Linux lacks:
Video card support
Wireless LAN support
"A good selection" of productivity tools.
OK, in other words everything needed for a flexible and high-powered server OS is right there. Notice how lame the "lack" list is, and that the same lacks also apply to the other Unixes -- even granting that there are widely-used video cards that XFree doesn't support (questionable, and how important is this in a server, anyway?), that X.25 doesn't count as a 'wireless LAN' (all the IT managers out there who depend on wireless LANs please raise your hands), and that "a good selection" requires more than two complete suites (StarOffice and Applix).
Thus adding additional video card support to XFree, additional kernel drivers for wireless LANs, and two or three more office suites will obviously bloat the Linux kernel, while SMP, multiple file systems and net protocols, and RAID support did not. Sure. Right.
This article would be more at home in a collection of postmodernist academic essays (or perhaps press releases from government bureaucrats) than in a technical journal; it's a wonderful example of content-free writing.
... under Linux & Netscape 4.08, default font either (TT) Verdana or (Adobe) Helvetica, 1280x1024. (TT font support for X at http://airnet.net/craig/linux and links from there) --
Slashdot Mirror
Well, it clearly didn't do it very thoroughly.... In fact, what it [business] did was break up the AI Lab at MIT in the midst of the "artificial intelligence" marketing craze of that era, the sole long-term effect of which was to offend RMS' communal sensibilities (understandably enough) and inspire him to dedicate his considerable hacking talents to e.g. gcc and other utilities central to OSS. Prolog and its friends are at best confined to a tiny niche; OSS is in the process of squashing NT, Microsoft's Prime Weapon for dominating enterprise computing. Some "crush."
But in spite of RMS' offended spirit, open software continued to be available from numerous other sources, both in academia (think about BSD and its associated stuff) and elsewhere. The "growth of the internet" by any reasonable metric started around 1983; basic free stuff -- like sendmail, vi, and so on -- were circulating from the very beginning. In 1991, when Linus first put up version 0.00whatever, the internet was a tiny fraction of its current size -- but it was big enough....
RMS himself, by the way, points out that nearly two-thirds of a standard Linux distribution consists of non-GPLed software, so apparently the FSF, for all its valuable contribution to OSS, is still in the minority when it comes to defining the term "free software."
Craig
I don't know anyone whose vision of life consists exclusively of anything. Do you?
Likewise I know very few people whose vision of life does not include food and shelter, though in some cases (notably RMS') in a rather minimalistic form.
> You assume that "enlarging" is a good thing. Why?
Gee, I dunno. I just looked around and noticed that a hell of a lot more great software with source was available now than it was when I first started using it nearly ten years ago. Silly me, I thought it might be because more people were writing the stuff....
> You also assume that some sort of plusralism in the motivations of a group is all right. Why?
I assume it's all right for the same reason I assume gravity is all right; it's inescapable. "A group" is not a thing, it's a word for a number of people who choose to cooperate with each other in some area of endeavor. Show me somewhere a group all of whose members are there with the same motivations, values, and goals and I'll show you a group that is at the very least unlikely to be individualistic enough to do any really superb hacking....
> Is it the freedom to change Linux into a system that relies upon utilities and applications that are not as Free as those distributed under the GPL?
The last time I looked, Linux depended for its GUI, for example, on X. Is that "more" free, "less" free, or "as" free as if X(Free!) were under the GPL?
Note that the definition of "free" promulgated by the GPL is oddly one-sided; many of us object to the habitual use of "free software" as a synonym for "GPLed software", just as we'd object to restricting the phrase "Let's do lunch" (now trademarked by Frito-Lay) to proposal of occasions when we planned on consuming potato chips.
Craig
-- as thousands of the talented do every year when they graduate and get a [gasp!] job!
> 2. siphon hackers out of the community as experts in Linux doing the bidding of the suits
-- that is, actually paying money to some hackers for working on Linux! Eeeek!
> 3. change the composition of the community by attracting those that are not interested in the primary goal of freedom.
-- that is, enlarging the community (Horrors!) by including people whose definition of "freedom" may not be exactly the same as RMS'. Aaagh!
Craig
Hmmm. The near-ubiquity of transistor radios in the Third World, the green revolution that showed up all the Malthusian doomsday scenarios of the '70s for the nonsense they were, the Ford Model T that set in motion the incredible mobility of the rural population, the ability of a quarter of a billion people or so all over the planet to communicate with each other instantaneously via the Internet, the revolution in "freedom of the press" brought on by the Web -- but these dreams are "seldom fulfilled." OK, sure.
If you're saying "Nothing is a panacea", that's true but trivially obvious. If you're saying any more than that, you're obviously quite wrong. Technological development in the last 150 years has vastly increased the empowerment of individuals.
Craig
Microsoft's habit of casually diddling the core of the operating system every time they need something new for marketing reasons has lead to fragmentation even within nominally identical platforms.
For example, if you apply Service Pack 1 to Visual C++ 5 and recompile your application, you discover that suddenly your code won't load on a machine running the original Win95 Retail. Why? Because SP1 changed a header file to default to using routines from a version of comctl32.dll -- one of Win32's most central libraries, the Common Controls library -- that was distributed only with Internet Explorer 4! You can fix your code once you've discovered the problem, of course, but documentation for this "upgraded feature" is totally buried in an obscure section of the documentation, and isn't even mentioned in the release notes!
Likewise many sites have found NT4 Service Pack 4 to break a number of crucial apps. So the "unity" of the Win32 platform is largely illusory anyway; at Microsoft, the perceived needs of the marketing department so far outweigh any technical considerations and user requirements that there's no contest.
And, when all is said and done, this is really what's hanging Microsoft. IT professionals all over the planet are sick of these marketing games and the resulting shoddy code; they're eager -- almost desperate -- for an alternative to Microsoft in the server room (which is where their most critical problems originate). And this is why they're leaping into Linux (and FreeBSD and even NetWare (again)) with glad cries and tears of joy and relief.
Craig
No, but why would you want to? The whole "shrink-wrap" ethos is basically an outgrowth of the PC explosion.
Note that game suppliers and the packagers of bundled software for (e.g.) scsi scanners simply provide both Win and Mac versions on the distribution CD -- and at this point differences between SCO, Linux, Solaris, AIX, and HP-UX sources are so well-known that operating from a common source base isn't at all difficult -- and the "intellectual property" concern leading to the distribution of binary-only code is based on the source, anyway.
One of the design mistakes made with OS/2 -- or at least one of the things that hampered its spread -- was that you couldn't simply recompile a Win application for it. IBM tried to remedy that, to a limited extent, with Merlin's compatibility APIs, but it was too late. Although you never know, it may yet recover in the current wave of interest in Microsoft alternatives; reports of its death have been greatly exaggerated before....
Craig
Wrong. You can release your executable any way you want, but if you make any changes to the library source they must be publicly available, a la GPL.
Libc is LGPL, for example, and every available Linux executable (including Applix, Netscape 4, StarOffice, etc.) is linked against it.
Craig
Someone posting illiterately as an AC who actually expects to be taken seriously when he accuses others of being "wannabes". How did you get in here, sonny? Didn't Rob card you? Or did you tell him you were looking for your daddy?
Craig
Yeah, and he does it pretty well. But keeping both processors busy on a desktop machine is a good trick; most of us are doing only one really heavy computing task at a time; downloading in the background hardly keeps the CPU awake, let alone busy.
But if you make -j3 your kernel on an SMP machine, for example, which tells make to spawn and manage three separate jobs, you'll notice quite a speedup; you'll save about a third of the time (not half, because of overhead) if you've got enough memory.
Craig
By the way, kfm's html handling is much improved in KDE 1.1; try browsing with it now. Some stuff is still not quite right, but it's mostly very usable.
Craig
This is odd; my impression of GNOME was that it was if anything slightly snappier than KDE. (Both built on my RH 5.1+ machine, egcs -O6 -mpentium.) How much memory do you have? How much else is going on in the background?
Also check your XFree version -- 3.3.3.1 is a little better with the Matrox card than 3.3.2.
Craig
http://airnet.net/craig/linux
Libjpeg is weird in that its header file contains both major and minor number together (61, 62, etc.) and it refuses to load without an exact match. The KDE devel crew uses .61 (I had this problem with GNOME, too, but they use .62 [sigh...].)
Try linking libjpeg.so.61 -> libjpeg.so.62, ldconfig, and see if it runs....
Alternatively: Get the kdesupport rpm, make sure libjpeg.so.61 installs in /opt/kde/lib, make sure /opt/kde/lib is mentioned in ld.so.conf, ldconfig, and see if it runs....
Otherwise just get the source RPM and rebuild it.
Craig
" Server busy or unable to fulfill request. The server is unable to fulfill your request due to extremely high traffic or an unexpected internal error. Please attempt your request again (if you are repeatedly unsuccessful you should notify the site administrator).
(Location Code: 25)"
Craig
... you have to configure imlib --with-GMODULE
or the piece of code libgnomeui needs doesn't get
built.
Also the intl/ directories in several of the
source tarballs are missing a header file; just
find it in one of the other tarballs and dump
it in...
Craig
If I'm right about this, it would seem to me that adding language bindings to a C++ library would be substantially trickier than to a C library.
Myself, I prefer to work in C++, but de gustibus non est disputandum (except of course that COBOL sucks and RPG is a joke...).
Craig
Over its history (and possibly still), GNOME stuff has lived in various places depending on when (0.2x vs now) and how (rpm vs tar.gz) it was installed. Does anybody have a script that will search and destroy all obsolete GNOME cruft everywhere?
Craig
still using KDE... and xfce, and wmx, and WM...
True but vacuous. Is there any general-purpose distro without plans to support 2.2.x, egcs, and glibc in the very near future?
The devil's in the details -- for example, the libjpeg problem with its unnecessarily fine-grained version code in its header. But these are easily solvable.
As for LSB, as long as it's a reasonable discussion between techies to solve technical problems, bless them. But -- at one point, at least -- it was in danger of becoming yet another vehicle for the overblown Perens ego, to which the various distributions replied (quite properly) by ignoring it and starting side discussions.
I have no idea why Love chose to write this now, unless there's some special situation at Caldera that requires a political statement of LSB support from them. I would have thought that the whole standards thing was mostly a non-issue by now....
Craig
$1000 hardware + $0 OS
selling to consumer for $1200
OEM earns $200 profit/machine
Craig
... just Katz' fatuously overblown rhetoric.
Craig
They were complete imbeciles, of course. So is Jon. But a little nostalgia every now and then is good for the soul....
Craig
> 95% market share means it's the best product.
Nope. 95% market share means only that for 95% of the purchasers, it's an acceptable balance between functionality and cost on the basis of their individual knowledge. It says nothing whatever about inherent technical quality, aesthetic appeal, efficiency, usability, or anything else.
(And since when, by the way, do us libertarians, who profess a radical individualism, believe that the majority is necessarily right? Politically we reject the tyranny of the majority as firmly as we reject the Divine Right of Kings.)
Now, if you (or any of us Linux -- or Be, or OS/2, or whatever -- advocates) think they're wrong, the answer is to educate them, and do whatever is necessary to raise the functionality and/or lower the cost (not just $, but hassle and time) of your favorite OS. And I think that's what we're doing -- isn't it? -- with improved installation, KDE/GNOME, etc.? (Who have you introduced to Linux today?)
Craig
See -- I'm one, too....
The heedless fascism of the GPL has caused so much grief to the rest of the free software community, though, that I'm somewhat amused to see the FSF having similar problems with a much freer license....
Craig
Look at all the contributors to the kernel mailing list on such topics as memory management and file systems (are these "CORE mechanics" enough for you?).
Everything is fascinating to somebody; Alan Cox got hooked originally because he became fascinated by, of all things, ethernet cards.
And believe me, if there's a fundamental bug in the memory management or your ethernet card driver, whether these things interest you or not you'll notice it, and probably report it in a whiny and offended tone!
Craig
If you actually analyze carefully Ted Lewis' article -- an exercise I don't recommend -- you'll see that none of his statistical arguments hold water -- even in terms of the mathematics, let alone their factual premises.
That is, where it isn't simply incoherent. Consider, for example, the following:
What does this sidebar actually say [p. 125]? That Linux has:- Multithreaded kernel
- SMP
- Multiple FS support
- Disk striping and mirroring
- Multiprotocol networking
- X11 GUI
- GNU tools
- all the internet tools
Linux lacks:- Video card support
- Wireless LAN support
- "A good selection" of productivity tools.
OK, in other words everything needed for a flexible and high-powered server OS is right there. Notice how lame the "lack" list is, and that the same lacks also apply to the other Unixes -- even granting that there are widely-used video cards that XFree doesn't support (questionable, and how important is this in a server, anyway?), that X.25 doesn't count as a 'wireless LAN' (all the IT managers out there who depend on wireless LANs please raise your hands), and that "a good selection" requires more than two complete suites (StarOffice and Applix).Thus adding additional video card support to XFree, additional kernel drivers for wireless LANs, and two or three more office suites will obviously bloat the Linux kernel, while SMP, multiple file systems and net protocols, and RAID support did not. Sure. Right.
This article would be more at home in a collection of postmodernist academic essays (or perhaps press releases from government bureaucrats) than in a technical journal; it's a wonderful example of content-free writing.
Craig
Craig