Re: The Problems With Linux

From: Rex Ballard (r.e.ballard_at_usa.net)
Date: 05/25/04 

Date: 25 May 2004 13:41:53 -0700

wintroll@yahoo.com (Suzie) wrote in message news:<ecc1c8e8.0405240952.3a718846@posting.google.com>...
> Here are the problems with Linux:

Yummy Yummy Wintroll - chomp chomp :-D

> o it is made by people (some of whom have beards)

Lincoln had a beard too. Beards didn't become unfashionable until
world war I, when it became necessary to wear gas masks on the
battlefield. Those who wore beards did not get a good seal, and died
- very painfully.

Many of those in the Navy still sported beards. Of course, Grace
Hopper, a Navel officer, was one of the pioneers of the computing
industry.

> who openly call themselves hackers;

Long before hacker became synonymous with those who committed cyber
crime, the hacker and hacker ethic was a badge of excellence.
Microsoft attempted to brand cyber criminals as "hackers" knowing that
this was the term used to refer to those who were very proficient
programmers of UNIX, Open Source software, and Shareware. To
Microsoft, programming for anything other than Microsoft's best
economic interest was a criminal activity. Bill Gates believed that
anyone who wrote software that diluted the value of Microsoft software
and products was stealing from Microsoft. In some cases, Microsoft
even attempted to file lawsuits against shareware programmers.

Ironically, one of the great contributions of "hackers" was their
commitment to improving the quality and security of code such as that
used in UNIX and Linux. Both systems were multi-user, multitasking
systems. Because of this, there was much more concern for the
security of teachers, who kept grades on the UNIX computers, and
administrators, who kept financial records on UNIX files and
databases, and system accounts, which managed the flow of
communications.

UNIX itself was a "hack". Brian Kernigan and Dennis Ritchie were
given some time and some cheap equipment, and cooked up UNIX as a
research project. AT&T could not "sell" UNIX because they were a
regulated monopoly and were barred from using their monopoly of the
telephone system to give them an unfair advantage over companies like
IBM, DEC, HP, Sperry, and numerous other computer companies who held
share in a fiercly competitive market.

Since there was no market for UNIX, AT&T donated UNIX source code to a
number of colleges and universities. Many of these universities and
colleges used UNIX as a teaching tool. Before long, thousands of
programs were written by these students, especially MSEE students, who
would work on very large and complex projects for their Masters or
Doctorial Projects. Often, these MSEE and PhD students supervised
undergraduates who participated at a more basic level, writing smaller
projects.

When AT&T finally accepted Divestature, they attempted to market an
AT&T "pure" version of UNIX - System III. The BSD system, which
included contributions from all over the country, had become so
popular that corporations would not purchase AT&T UNIX without the
programs only available from BSD. BSD had also made hundreds of
enhancements to the kernel, especially those used for VAX systems,
including demand paged virtual memory, client/server technology,
interprocess communications (sockets), and most of the TCP/IP
technology.

AT&T traded nonexclusive rights with BSD. AT&T had nonexclusive
rights to BSD code, and BSD had nonexclusive rights to nearly all of
the AT&T code. Each system was published under their respective
licenses, with the permission of the other.

The AT&T license gave those purchasing UNIX from AT&T the obligation
to pay AT&T royalties, and to limit their disclosures of the source
code and related techonologies.

The BSD license gave those purchasing UNIX from BSD the permission to
distribute most of the technology in source code form, and even the
ability to market the BSD code as proprietary products. For example,
BSD code is used by Microsoft in it's Windows NT/9x/2K/XP code. The
only requirement of the BSD license is that BSD be included in the
credits.

> o it involves stolen copyrighted code from a highly reputable company;

The only "stealing" going on is the attempts by Daryl McBride to
"steal" code which had been legally traded to those holding the UNIX
copyright over the last 30 years. McBride seems determined to attempt
to enforce a contract signed back in 1983, BEFORE AT&T assigned UNIX
rights to BSD, BEFORE AT&T traded rights to interprocess
communications, BEFORE AT&T traded rights with IBM, HP, Sun, and BSD
in exchange for NONexclusive distribution rights to X11, project
Athena, and numerous other open source tools.

McBride wasn't there. He has a very selective memory, and can't seem
to remember that Novell assigned nearly all of the rights to UNIX - on
a nonexclusive basis, to BSD - including the FreeBSD project, on a
royalty free basis. He seems to gave forgotten that all of that BSD
licensed code could legally be used on Open Source systems, so long as
the attribution, warranty, and related license terms were accepted and
implemented.

McBride has a history of nearly criminal behavior, and unethical
behavior throughout his past. But this should be disclosed by someone
other than me.

> o the priorly-assumed creator of it admitted this week
> that he didn't in fact create it,

Linus wrote a 10,000 line (source code) kernel, which was so different
from anything previously written that it would be almost impossible to
claim that it was NOT an original work.

Almost immediately, he published it under the terms of the GNU Public
License. Those contributing clearly understood the terms of this
license. Libraries were published under the LGPL, and many
applications were published under their original licenses (including
BSD).

There is a System V "compatibility library" for Linux, which was based
on the BSD documentation, and was published under the LGPL.
Applications written for System V UNIX (such as SCO's) could be run
under Linux, but the implementation was significantly different from
the AT&T libraries. At the time that these libraries were created,
APIs could not be protected by copyright or patent law. Furthermore,
the BSD license expressly permitted disclosure.

> but was posing as front man for the real creators;

Linus published the source code to Linux under the GPL. Those who
contributed enhancements, did so with the full knowledge of the terms
of the GPL. Perhaps Daryl McBride might be able to find one of those
contributors who is willing to confess to criminal charges of willful
copyright violation, go to federal prison for multiple years, and
claim that they wrote critical portions of the kernel. Since these
contributions are very carefully tracked (for just these reasons), it
should be fairly easy to identify exactly which code was illegally
contributed, and whether an alternative would be available.

Of course, there is always the that other little problem, that any
attempt to enforce such a violation might be contrued as a violation
of the terms of previous intellectual property agreements. For
example, SCO could find itself with no graphics, no communications, no
virtual memory, no security patches, no availability patches. In
fact, they might find themselves only able to legally market System
III in it's original form. They couldn't even compile it for an Intel
chip, since most of the Intel code come from BSD, FreeBSD, SunOS, and
Linux.

Perhaps they can market OS/9.

> o it is for geeks;

This would be like saying that any car other than a Ford Escort is for
"hot rodders". Microsoft has maintained their monopoly by adhering to
a "one size fits all" solution for decades. For a brief period, they
offered both Windows NT 4.0 and Windows 9x. Ironically, nearly all of
the Windows NT/2K offerings were primarily to counter the threat of
Linux/UNIX. When NT was announced, Dell was selling SCO UNIX on their
PCs, Sun had captured 35% of the corporate desktop market, especially
in financial services.

Until Windows ME, Microsoft's REAL product was the basic Windows 3.1,
95,98 system. Windows ME was so bad, and offered so little advantage
that customers were refusing to purchase systems unless the dealer
removed Windows ME and installed Windows 98. Many dealers had to give
customers free copies of Windows 98 "downgrade". Ironically, many of
those who purchased Windows XP prior to Service Pack 1 - insisted on
"downgrade" capability. OEMs had to write special drivers to make
sure that their boxes could run Windows 98 or Windows 2000.

> o it is too hard for most people to understand; and

So is a fuel injected car with an automatic transmission. Linux
offers the unique ability among computer systems to have someone other
than the original publisher (Microsoft) actually diagnose and correct
technical problems. Those who purchase a Rolls Royce can only have
their car serviced by a Rolls Royce mechanic. The hood is locked
shut, and only those who work for Rolls Royce have the key.

Chevy makes a number of cars which can be fixed using box-end wrenches
and screwdrivers. Mechanics anywhere in the country can fix a Camero.

Linux is more like that Camero. Corporate users can hire their own
fleet mechanics to provide server support and desk-side support.
Retail customer can purchase online support or can purchase desk-side
support from a Linux-friendly dealer.

The big problem that most new Linux users have is that they attempt to
download an unsupported version, over a high speed link, then try to
install it themselves, and because they didn't get a supported
version, cannot call anyone - such as the vendor, to get support.

Linux is free (as in beer), but support isn't. If you want to
download "version of the day" bugs and all, you can do it for free.
If you want a version that has been tested with commercial
applications, on commercial hardware, using the latest peripherals,
cards, and options, then you probably need to pay a bit more. SuSE
Professional costs nearly $100 but works on nearly every machine made
by IBM, HP, or DELL. Red Hat Enterprise Workstation costs nearly
$180, but includes much better support.

What really matters is the bottom line. Those who have replaced
Windows NT/2K servers with Linux servers have realized savings of up
to 90%. Put another way, the real costs, over the life of the typical
project, can be as little as 1/10th the cost of a comparable Windows
based system. The industry average savings seems to range between 1/3
to 1/4 the cost of NT, depending on the application and system being
replaced. Again, these are "real world" applications/systems rather
than just a carefully targeted single component such as a single
processor web server.

> o running programs in it takes too much effort: we want to double-click!

Writing scripts does take time. The good news is that you don't have
to do the writing yourself. Linux can do everything Windows does -
with a click (Linux applications are started with a single click by
default). If you just want the standard tools, such as word
processing, spread***, presentation graphics, e-mail, web browser,
and chat, Linux has that. You even have choices. You can use
KOffice, or get better MS compatibility with OpenOffice, or get really
good import/export compatibility with MS Office with StarOffice.

Again, the user can make the choice between different degrees of
quality/economy. Microsoft likes to charge BMW prices for a Ford
Escort, Linux offers an economy, mid-size, and luxury version,
depending on what you are willing to pay in terms of both hardware and
software costs.

> Bye

<Gulp>