Linux no threat to Microsoft

===Linux is no threat to Microsoft===
* - Scott Nudds - August 10, 2006 -*

It's been about 15 years since the original version of the Linux kernel was
produced by Linus Torvalds, and over those 15 years, Linux has grown from a
toy into a functional OS with some potential. My congratulations to all
involved.

The maturation of Linux has been reasonably swift - due essentially to the
duplication through the direct copying or reverse engineering of all manner
of features/tools/utilities found in the Unix environment. So although
reasonably swift, still probably encompassing several hundred thousand man
of development, the vast bulk of Linux development is a result of the
million or so of man years of development of the various versions of the
Unix operating system on which it has piggybacked.

The Linux desktop(s), (probably the most impressive feature of the OS), are
equally dependent on the open source available for X-Windows from which they
were similarly derived.

5 years ago, Microsoft could see Linux over it's sholder, rapidly
approaching and in their view a potential competitor on the desktop and
elsewhere. At that time, it was said that Linux commanded 2% of the server
market, and about 2% of the desktop market, and was thought to be growing.
Today Linux occupies about 25% of the server market but still retains only
about 2% of the desktop market or perhaps a bit less.

*Why hasn't growth of Linux been equally strong in both the server market
and in the desktop market?*

The answer is simple. Linux, like Unix, has always been, and remains user
hostile, and unsuitable for the desktop market.

The server market is different because the server market is maintained by
trained experts who largely enjoy the challenge of dealing with complexity.
The desktop market, on the other hand, consists of people trying to get real
work done.

Now lets be clear, SIGNIFICANT improvements have been made in the way Linux
operates over the last decade, and the GUI has matured to the point where it
is almost as good as the GUI provided by Windows 2000 and XP. Yet Linux
continues to be rejected by 98% of desktop users and 75% of those
implementing web servers.

*Why?*

This answer is also clear. Unix has always been an extremely user hostile
OS. The fundamental design philosophy behind Unix and therefore Linux has
always been extremely flawed, and hopelessly inferior to the design
philosophy of Apple - later adopted by Microsoft. The Unix and by
derivation the Linux community embody with a religious zeal, a loving desire
and passion for misplaced minimalism and that which is the purposely
cryptic, puerile, and adolescent.

As an example, while looking over a random snippit from a Linux driver file
the other day I found a setting called "HappyMeal". Professional
programmers do not engage in such puerile behavior.

In the early days of Microsoft, when machine cycles were relatively scarce,
and time was measured in microseconds, not the picoseconds measure that is
used today, Microsoft too used relatively cryptic commands typed from a
keyboard to command the OS. But while the Unix computing universe was busy
suing itself into near oblivion and issuing commands like grep, yak, ls,
chron, and building tarballs, Microsoft was busy listing directories,
finding files, and building file archives.

Microsoft won the war for the desktop, even when the desktop was a command
line interface because in part, Microsoft recognized that nomenclature
matters. Ease of use matters. Clarity of design matters. Consistency
matters, and yes, documentation matters.

These are lessons that to this day, virtually <ALL> Linux/Unix developers
have failed to recognize.

Consider this quote taken from the WIKI for blender - an open source
graphics development environment -

*"Users will not sacrifice efficiency for an intuitive UI".*

Excuse me... GUI's are nothing but a sacrifice of efficiency for the sake
of an intuitive UI. Blender exists for the express purpose of allowing
people to waste time drawing pretty pictures and performing unnecessary
animation - hence increasing the overall inefficiency of the human species.
Unix has only become a small blip on the radar screen of computer users
<BECAUSE> an inefficient GUI has been developed for it. Where would the
internet be without the inefficient UI that is provided by Graphical Web
Browsers?

Despite all evidence to the contrary, and all common sense, the above quote
actually encapsulates the land of delusion that virtually all members of the
Linux faith reside in.


*Not all of the Linux/Unix faithful are hopelessly clueless.*

It may be argued that since the early days of X-Windows, Linux/Unix users
have recognized these things, and have responded, but the truth is, for the
most part, X-Windows is primarily used to provide multiple command line
terminals into the Unix operating system. With Linux however, the GUI
interface is more pervasive and more powerful, and yes, proof that SOME
small minority of Linux developers have actually come to understand what
Microsoft (and Apple) knew 20 years ago.

As good as KDE and GNOME are, they are in large part, used to provide a
graphical interface to a text based, command line driven function that sits
below. Unlike Windows applications, the GUI in Linux/Unix is not integrated
into the application because doing so would violate the Unix/Linux design
philosophy. The result is a GUI that is designed to be networked, or more
precisely a GUI that is designed to interact only sparsely with an
application in a manner that allows the GUI to be run on a remote terminal
while the core code of the application runs on some central server.

As a result the X-Windows/KDE/GNOME desktop performs sluggishly compared to
the Windows GUI, and is incapable of the kind of fine grained interaction,
speed, and efficiency that you can get via windows. If the juxtaposition
of the words "efficiency" and "Windows" has made you laugh, considering the
dozen or so layers in the KDE/GNOME GUI interface should make you cry.

5 years ago, Microsoft saw Linux in it's rearview mirror and saw it is a
potential threat. 5 years later, Microsoft has developed DotNet programming
paradigm, and now it has a completely new VISTA API set to stave off the
Linux challenge. During this time, Linux has matured, but at the same time
stagnated at a point where it can no longer offer a serious challenge to
Microsoft.

Indeed, not only has Linux stagnated, but has actually declined in
functionality. I have been told by several people that Mandrake Linux was
significantly more functional and manageable, than current versions of
Linux.

5 years ago, Linux was not ready for prime time. Today Linux is still not
ready for prime time, although closer. Windows on the other hand has been
prime time in my opinion since 95 and with Vista and with DotNet Microsoft
has created an un-cloneable operating environment through a massive
expansion of it's API.

The hope of Linux was that it could provide interoperability with Windows
through the Wine emulation layer, support for NTFS, MACos, etc. This is no
longer possible with Vista of course. Proprietary interfaces can be
developed faster than they can be reverse engineered. Where is Linux NTFS
support? Further, Wine has taken 15 years to develop, and the new Vista
interfaces are large and complex enough that it will take considerably
longer to clone. Where will Windows be in 30 to 40 years when viable
clones of Vista might be coming available?

Compared to Microsoft Windows, Linux still remains largely user hostile,
although within the last couple of years, not necessarily terminally so.
Different distributions of Linux for the most part provide their own
graphical interfaces to custom or standard text based system utilities, so
reasonable progress has been made. Unfortunately many of these utilities
offer only partial functionality, inadequate functionality or no
functionality whatsoever. And of course there is the Gnome vs KDE
interoperability fiasco.

Is the Linux community best served with multiple, incompatible GUI's? Is
the Linux community best served by having to worry about which version of
Program X to install (or write) because GUI-A is incompatible with GUI-B.
If the answer is yes, then the Linux community should be even better served
with the development of yet another incompatible GUI, and another, and
another, and another, write a million of them. It's good for the linix
community.

If on the other hand the answer is no, then the Linux community will be
served best by one GUI. Why not choose one?

Is it an efficient use of time and effort maintaining and developing two
independent and incompatible GUI's and in many cases two different versions
of a program for each version? How much development effort is lost to this
kind of inefficient duplication of effort? How many extra bugs are created?
How much more debugging time is needed? Meanwhile, Microsoft with it's one
core GUI-API, does not suffer from this problem.

Is the Linux community best served with multiple software installation
utilities? CNR, RPM, Yum, Apt-Get (shouldn't that be "Get_Application"? by
the way? Too complex a concept for you guys?) Is the community best served
by 5 of these utilities? If so, then why not 6, 12, 100? 1,000,000?

How about just one. One that works. One that shows me what is being
installed and why?

Oh, by the way CNR = Click aNd Run. How about CAN = Click And Run? How
about "Remote_Install" How about "Get_Remote_Application"? How about
"Install_Remote_Program"? How about "Install"? How about
"Install_From_Network"? Or how about "Install_Via_Network"? Or
"Install_Via_Repository" Or "Install_From_Repository"? Hell, I'd even take
"Get_Remote_App"....

CNR = Intolerable ineptitude.

I have been casually auditing many of the current (mid 2006) distributions
of linux, Redhat, Ubuntu, Suse, Linspire, and a couple of others who's names
escape me at the moment, and with the exception of Linspire, <ALL> have been
erased from existence within a few days of installation as a result of
intolerable user hostility, inability to support standard hardware or
intolerable configuration issues that can't be solved without dropping to
the command line and editing impossible to find config files, documented by
incomplete text files that were written by an adolescent for whom English
was probably a third or forth language.

With the exception of Linspire, no version of Linux was capable of
multimedia output on a fresh install. Networking typically installed
reasonably, but connection to local Microsoft networks did not, with
Linspire again being a notable exception. Support for reading/writing NTFS
partitions is nowhere, and typically setting a PC up for duel boot with
Linux/Windows was a harrowing experience simply due to the fact that by
default, Linux installs typically wipe the entire drive clean.

*You can't get much more User hostile than that.*

Custom installs failed on two of the versions of Linux, on one with the
custom partition creation failing because custom partitioning was ultimately
not supported by the very installer that invokes it.

One OS died with a kernel panic after installing an audio driver, and one
installation failed to provide a mouse cursor. Ultimately a cursor was
turned on via software emulation, this version lasted a few days until it
became obvious that it had several troubling incompatibilities with several
pieces of standard hardware in the machine.

Currently my network card is configured at 10mb/s even though it's a 100mb/s
card. No facilities other than a text editor are available to change the
configuration whereever it is. Additionally the sound card mixer control is
buggy, randomly confusing mixer channels and providing a very non-linear
response to the graphical state of the various volume sliders. Sometimes
adjusting the volume causes one channel to simply cut out.

Ubuntu would run from CD, and would install, but then failed to provide a
graphical desktop due to it's inability to understand the Nforce chipset.
This problem may not exist on the latest version of Ubuntu, but it's lack of
a admin account and lack of multimedia support and lack of connectivity
support to existing windows machines precipitated it's erasure.

One particularly galling "feature" of Linux is it's lack of device
management support. None of the versions of Linux that I have audited have
provided anything close to functionality of windows device manager for
managing and altering the performance of hardware devices. I consider
Windows device manager to be only marginally adequate. Marginally primarily
due to the lack of a big red hardware device reset button for each and every
device.

*Linux on the other hand has nothing.*

*Correction*, several Linux versions had "device managers", that reported
device settings, but no editing capability was provided, and hence no
ability for the "device manager" to manage anything. Certainly not my anger
at the dishonesty of including the word "manager" in the title of something
that does not, and can not manage anything. Maybe the title reflects some
wishful thinking.

I can manage to be insulted like this by Linux only for short periods of
time, perhaps an hour, if it's hot, perhaps a half hour, and then the OS is
shut down in disgust.

Earlier today I decided to check out OpenOffice Math. Oh, what is this? I
ask myself. Looks like an equation solver, perhaps it's mathematical like.
After looking over the documentation for 10 minutes do I find anywhere -
ANYWHERE - an introductory paragraph telling me what this application does?
No. Of course not, in Unix/Linux, the philosophy remains to offer maximum
hostility toward the user. 10 seconds on the web, and poof, a one paragraph
overview of the program.

Time to shutdown the OS and contemplate how it is possible for those who
have written and proofed the documentation, and not seen a need for an
introductory overview of the program can manage the brain power needed to
breathe and sit upright at the same time.

How old is OpenOffice? 10 years? Why after 10 years does the documentation
start with a description of how to link some meaningless widget to some
meaningless file, when it any thinking person would realize that the
documentation should start with a short description of what the program is,
what it does, why it exists, and how it works?

Hi, my name is Bob, and I'm your waiter for tonight. Here is a menu. I
recommend the pasta Elfredo this evening.

Resteraunt servers seem to manage these concepts quite nicely. But not the
members of the Unix/Linux faith. It's some form of collective brain damage.
User hostility is their stock and trade.

I am trained in the sciences, and have been programming computers for almost
30 years, largely machine code initially, so I am no stranger to complexity
and the cryptic. I am also no stranger to logic and reason. In the
sciences we are trained to write papers with the following sections,
Purpose, Method, Apparatus, Observations, Conclusions, Data, and finally
References. This represents a logical progression from initial idea,
through the testing and verification or disqualification of that idea,
presented in a form that can be honestly replicated by others.

The C/C++ programming community and by extension the Unix/Linux community
appear entirely immune to these rational constructs, excluding purpose, and
method, only partly describing the apparatus, and then documenting their
conclusions in pig latin, with the proclamation that the observation section
is unnecessary because the data is self documenting. A philosophy of
ignorance, unprofessionalism, and maximum user hostility.

To be fair, can Windows be frustrating, and it's documentation poor?
Absolutely. Does windows help actually help with anything? Not in my
experience. Yet Microsoft documentation continues to be superior to the Man
pages provided for Linux. And even if it can be argued that they are
roughly equivalent, why strive to recreate failure? Shouldn't the goal be
to exceed Microsoft? Shouldn't the goal be to excel and progress?

5 years ago, Microsoft looked over it's sholder and saw Linux gaining
rapidly. 5 years ago, Linux had 2% of the desktop market. Today Linux has
2% of the desktop market if not a little less. In those 5 years, Microsoft
has developed the massive Vista API and the massive push to the DotNet
programming paradigm of just in time cross assembly. Linux on the other
hand has matured but remains not ready for installation on the desktop in
anything but turnkey systems where the user is presented with everthing
preinstalled and preconfigured to work out of the box, on a PC who's
hardware has been tested for Linux driver compatibility.

Linux in my view is in the state that makes it roughly equivalent to a
windows 95 beta, more complex and mature than Windows 3.3 but still not
ready for general release, still relying heavily on DOS (command line Unix),
and not quite compatible even with itself.

Linux has in fact stagnated. There is no hope to clone the Vista API, and
while Windows continues to advance in this direction, Linux can at best hope
to maintain the ability to emulate what is now the outdated Win32 API.
Linux can only piggyback on the success of Windows for a few more short
years before it's effective compatibility with windows will be gone. At
that point it will have to fly on it's own, and on it's own merit.

Are three to four years enough to allow Linux to grow enough to challenge
Vista? No. Even with all of the resources available to Microsoft, Vista
has taken a half decade to develop. It is un-cloneable, unemulateable, and
developing similar infrastructure for Linux will take significantly longer
if it's even attempted.

Provided Vista is a success - and this is a legitimate question - Linux has
no hope of challenging Microsoft on the desktop for the forseeable future.
Linux will therefore remain a small player in a market dominated by
Microsoft operating systems. At this point Linux can provide no serious
challenge to windows.

Now on the other hand, should Vista fail, Linux might make some headway as
Microsoft retools and reblends the Vista components into something more
palletable. But make no mistake, the Vista API will persist, for the sake of
backward compatibility alone. Repositioning and rebranding Vista may take
Microsoft a couple of years, but in those years Linux can only hope to reach
the W95 level of maturity before Vista 2.0 becomes a reality. So at best,
Linux might gain some ground, perhaps a percentage point or two. 100%
growth over 2 years would be quite unprecedented and explosive, and hence
growth to under 4 percent of the desktop market is remote even under the
best of circumstances.

Microsoft therefore, has nothing to worry about with Linux. Contrary to the
statements it made 5 years ago, regarding the challenge Linux posed. (Lets
face it, those comments were primarily made to strengthen it's hand against
the failed drive to split the company). Even then Linux posed no probable
threat.

Linux will for the forseeable future remain user hostile and simply wrong
headed in many respects, (case sensitive string comparisons immediately come
to mind), while Microsoft continues to churn out improvements and real
advancements.

At this point the only real threat to Microsoft on the desktop, is Microsoft
itself. *Vista might not fly*.

The continuing migration toward Microsoft oversite, control, of both the OS
and the computing environment - particularly when it comes to DRM -, might
blow a big hole in Microsoft's foot and operating budget. We will soon find
out if Balmer's Vista vision is Microsoft's new coke tragedy, or if the
American sheeple are as dumb as the rest of the world thinks they are and
will swallow the Vista stinker along with the Vista hook. It will be
interesting to watch.

You might wonder why am I auditing Linux when I obviously feel the OS is
fundamentally flawed, and rotten down to (but not including) the kernel/GUI?

This is because, Vista is going places that I don't care to go today,
tomorrow, or ever. I prefer computing freedom over corporate data
monopolies, and freedom of speech over corporate gag orders. I prefer user
oversite over buggy concealed code. I prefer the hope of user
serviceability over that which is proprietary. I would prefer my computing
future not being mandated by corporate interests and artificially controlled
product lifetimes. So I will happily jump the Microsoft ship, (which I have
never supported), in favor of any OS that provides the same or better ease
of use and hardware/software interoperability that Microsoft Windows (2000)
currently provides.

Unix and it's derivatives however represent a distasteful choice not only
because they fail the primary criterion for change, but also because of the
history of ineptitude shown by Unix vendors in the early 80's as they sued
each other into the ground, fighting over an ever shrinking wedge of pie
that was so clearly being eaten by Microsoft and it's inferior products.

Unix provided no coherent challenge, no linspired vision, no set direction,
no design Plan9, no managed strategy, and therefore no threat to Microsoft.
Historically Unix vendors have retreated at every market challenge presented
by Microsoft. They were and remain to this day, functionally adolescent,
unprofessional, cowardly, disorganized, largely brain dead, and
frustratingly impotent. Today as yesterday, the Unix/Linux community still
can't manage the collective brain power to realize why Microsoft continues
to clean their jiffy-clocks.

Microsoft continues to be dominant because it knows that ease of use
increases productivity, ease of use decreases error and and ease of use just
makes computing more enjoyable than the insulting Linux/Unix alternative of
memorizing cryptic nonsense found in random config files, or poorly named
switches associated with childishly named functions.

*I and a circle of friends, collectively represent about 200 years worth of
computing experience, and we are all looking to jump from the bloated
Microsoft Windows platform, and have been for years. Yet we are all
absolutely disgusted by the equally bloated, still buggy and incomplete
state of the Linux/Unix environment. Linux/Unix is today as Unix has always
been, user hostile and inferior, which is a shame, and a shame that is
directly attributable to the underlying design philosophy of C/C++. and by
immediate extension Unix/Linux.*



.