Linux-Advocacy Digest #896, Volume #25 Fri, 31 Mar 00 23:13:04 EST
Contents:
Re: Why did we even need NT in the first place? (R.E.Ballard ( Rex Ballard ))
Re: Nice link (Andy Newman)
Re: Top 10 reasons why Linux sux (Daniel O'Nolan)
Windows 2000 has "issues" (Robert Morelli)
----------------------------------------------------------------------------
From: R.E.Ballard ( Rex Ballard ) <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Why did we even need NT in the first place?
Date: Sat, 01 Apr 2000 03:18:21 GMT
In article <[EMAIL PROTECTED]>,
[EMAIL PROTECTED] (mr_rupert) wrote:
>
> Can anyone remind me why the computing world
> needed a new server OS?
Keep in mind that the original driving force for NT was that
Sun was getting some very strategic positioning as a Corporate
desktep in 1990 and 1991. Windows 3.0 was cute, but it crashed
regularly, had very limited multitasking, could only run one
MS-DOS application at a time, and required an $8000 computer
to run on. The Sun SLC ran multiple concurrent windows, for
weeks without as much as a hiccup, supported Lotus 1-2-3, Applix,
FrameMaker, and many other top of the line applications that
had been locked out of the Microsoft Windows 3.1 market.
At the Winter CES, Sun was getting all the attention and the
Windows 3.1 had already become a laughing stock in the community.
Many vendors weren't even showing Windows 3.1 applications, and
severel analysts were predicting that UNIX would kill Windows
in about 3-5 years. Bill Gates had been scheduled to give the
keynote speech on the closing night. After seeing the attention
that the Sun machine had been getting, he and his top techies
began studying the crowd's reactions. They listened for what
the people liked, and what they didn't like. When it was Bill's
turn to speak, he annouced NT. I doubt that you would find even
one byte of code for NT prior to that speech, but Bill hit all
the right buttons. He promised real multitasking, multilayered
windows, and better reliability through better memory management
and better use of the 32 bit addressing of the 80386 and 80486
chips.
For the next three years, Microsoft was in a race for it's life.
They spend billions of dollars to advertise in publications that
give favorable reviews to previews of NT, to Windows, and to
Microsoft applications - and ignored any variation of UNIX.
Especially this little upstart called Linux. One of the most
heavily punished was Byte Magazine, who had helped create Microsoft,
and was now being starved by Microsoft's control of it's own
advertising, AND the Advertising of most of the OEMs and VARs,
through CO-OP money (a portion of the OEM add paid by Microsoft).
Even Dow Jones was punished several times by the cancellation
of full-page-ads when they gave favorable reviews of UNIX. The
pressure got especially intense when Novell purchased the rights
to UNIX, and appeared to be gearing up to create a UNIX based
Workstation. The UNIX community was annoyed that Novell was selling
UNIX for $200 and selling TCP/IP for $400, but the key was that
Novell was in the marketplace.
Microsoft really hadn't planned to use NT as a server. They even
had a signed agreement with Novell that stated that they would keep
NT out of the Server market if Novell stayed out of the Workstation
market. This agreement upset Ray Noorda so much that he resigned
his post as CEO of Novell and left to back Caldera.
It wasn't until about 2 weeks after the official public release of
Windows NT that Microsoft realized that they had a very serious
problem. When NT 3.x was released, it required nearly twice as much
memory as Linux or UNIX, nearly three times the processor speed,
and nearly 4 times the hard drive space.
Microsoft had hoped that hobbyists and computer techies would
snap up a few million copies during the first month. They were
hoping that these technophiles would then proceed to convince
Management that everyone should upgrade to Windows NT. The
stratagy backfired. The technophiles who knew anything about
UNIX were suggesting that the corporate workstations should be
switched to UNIX since the workstation and the server could team
more effectively.
Microsoft was able to persuade a few corporations to commit to
making the switch. Many of these companies held a substantial
equity stake in Microsoft and hoped that their glowing endorsement
of NT would cause the rest of the market to flood in. Again,
it backfired. Horror stories of companies spending $10,000 per
user to back up the old system, install new hardware (usually
complete replacement of the box) and then restore the user's data
were commonplace. What was worse was that many key strategic
corporate applications wouldn't run on the new NT machines, and
some corporations were actually suffering badly.
What was worse was that UNIX had come in the back door - via the
web browser and was effectively dominating the application
programmer space. Many institutions had discovered that while
custom programmed applications sucked up resources and cause
Windows 3.1 to crash regularly, you could run a free web browser
or very cheap Netscape browser on Windows 3.1 and let a UNIX box
do all the work. Microsoft was loosing mind-share to wannabe
web-masters who were loading up Linux to learn UNIX and Internet
technology as fast as they could.
It took about 6 weeks for the situation to really sink in. Bill
Gates had spent several $ billion on what appeared to be an investor's
nightmere. Not only wasn't NT selling, but with the market threatening
to go UNIX as part of the Internet craze, there was a very good
chance that Microsoft would either loose control of the market
entirely, or they would loose the profitibility from the software
sold to OEMs. Meanwhile, the OEMs were getting a bit upset because
the new NT PCs weren't even being ordered.
It was right about this time that Bill Gates made two key
announcements. First, Microsoft would offer an NT server.
This assured that Microsoft could sell some boxes, and it
kept the OEM from defecting to one of the UNIX vendors. Dell
had actually started selling SCO Unix machines.
The other key announcement was "Chicago". This was supposed to
be a new version of NT that would run on existing 80386 and 80486
machines, allowing corporations to upgrade to the new operating
system without having to replace everything. Initially, it was
supposed to be out within about 6 months, hopefully around the
first Quarter of 1994. The goal was to get Chicago out in time
to cash in on the christmas market and the tax season, and
be included in 1994 fiscal budgets.
Unfortunately, there were delays. The first problem was that
people were getting tired of the Windows 3.1 interface. What
they wanted most was the ability to switch between applications
when the forground application was showing the hourglass. The
HourGlass had become the topic of many jokes. The overlapping
windows was also a problem and Microsoft wanted to encourage users
to run applications in full screen mode, switching from application
to application by pressing buttons at the bottom of the screen.
The program manager was also a problem because it was often getting
covered by applications that went into general protection faults
or when a background window tried to pop up an OK/Cancel dialogue.
They added the start button at the bottom of the screen, along
with a "shut down" selection that allowed the user to kill applications
that locked up due to OK/Cancel dialogues or gpfs.
By about September of 1994, it looked like Chicago might be
coming out in time for Christmas. But Bill Gates saw the biggest
threat to hit Microsoft since the Mac. A little company called
Yddragasil had come up with a version of Linux called "Plug and Play"
Linux. Not only was this the easiest installation of UNIX available,
but it was also easier to install than Windows 3.1 or Windows NT.
It could pretty effectively configure an SVGA compatible card, an
ISA, EISA, or VLB motherboard and the most commonly used peripherals
(keyboard, mouse, printer, modem, and video) and the installation
took about 20-40 minutes depending on the speed of the CD-ROM.
Byte, and several other magazines that had been snubbed by Microsoft
for covering Novell UNIX and OS/2 were more than happy to give
Yddragisil very generous reviews. Microsoft immediately set it
technology team in motion to create it's own version of "PlugNPlay"
(talk about a trademark violation...). More important, Microsoft's
legal team began some of the most aggressive arm-twisting in
Microsoft's history. Microsoft wanted to be absolutely certain
that Linux could not implement "Plug and Play" on Microsoft's
"Plug-n-Play" technology. Every OEM, card vendor, chip maker,
and peripheral maker was told to use the PCI and was told to sign
a very comprehensive nondisclosure agreement that was specifically
designed to prevent any information AND any reverse engineering
by the UNIX community or the Linux community.
The new "Plug-n-Play" delayed Chicago almost another full year.
Microsoft had announced that this new product would be called
Windows 95, and by August, people were wondering if that meant
it would be released in 2095. On the very last saturday of
Microsoft's 1995 fiscal year, Microsoft officially released
Windows 95. To fluff up the annual report, Microsoft actually
financed the "sales" of Windows 95 directly to the retailers,
and financed orders for some of it's corporate customers. On
paper, it looked like Microsoft had sold 1 million copies of
Windows 95 on the first day. In reality, many customers
shied away when they realized that along with their Windows 95
upgrade kit, they would also need to upgrade their 4meg RAM
to 16 meg (about $500-$700 at the time). They also needed
to upgrade their 40 megabyte hard drive to at least 200 meg
(another $400-$500 at the time). Many people looked at that
$1000 "upgrade" and literally left the stack sitting on the
counter. In New York, I watched as the CompUSA sales people
went from elation to depression. Elation in the morning
because they were certain that everybody would be spending
$700-$1000 a pop to upgrade to Windows 95. Depression in
the evening because nearly all of the inventory (which they
had purchased in advance) was still sitting on the shelf,
along with shelves of overstock hard drives and boxes of
overstock DRAM.
Actually, by the end of the week, there were people standing outside
of the store telling people not to try to upgrade to 95 because
the PNP didn't work on ISA and VLB machines.
The OEMs actually hit the jackpot - with the exception of IBM, of
course. Since users really needed the PCI cards, and they needed
new hardware for the upgrade anyway, there was a big push in the
corporate community who had now been waiting almost 6 years for
the "Better UNIX than UNIX" to arrive. Many of them had literally
bet their jobs on this new technology, and many were looking at
unemployment if things didn't work out.
Ironically, many of those who had bet the farm on NT were still
under the sword of damacles, because they had purchased NT servers
in hopes of replacing UNIX servers and protecting Microsoft's
interest in the desktop. Unfortunately, the NT 3.51 servers
were still only able to serve about 10-30 users depending on
the processor size,the amount of RAM, and the complexity of
the application. Many CIOs, VPs, and CTOs were still standing
with egg on their face.
Worse, Red Hat had obtained the PnP technology from Adaptec who
felt that Microsoft had not properly managed the relationship
(Just as IDE had maxed out, Microsoft came out with FAT32, LBA,
and EIDE). Worse yet, InfoWorld was singing the praises of
Linux, especially Red Hat, and wasn't being bashful about it.
Worse, Linux was being compared to Windows 95 and Linux was
winning. Linux was faster, more reliable, and even had a nearly
identical user interface (FVWM95).
Microsoft went into NT 4.0 development with a vengence. Microsoft
knew that if Linux made it to BOTH the server market AND the desktop,
that they would lose control of the market. Microsoft adopted the
superiour 95-style Window Manager to NT, changed the software
architectuer to reduce the kernel vulnerability to application
corruptions, and reworked the scheduler and the memory manager.
Microsoft knew that they had to get it right. If Windows NT 4.0
wasn't at least good enough to be compared to UNIX, that they
would lose control of the mind-share. To get back the developer
base, they started their MCSE program, and they tried to tout
NT as a distributed processing machine. Windows NT 4.0 was much
better than it's predecessor, it could easily support 100 users
when equipped with a Pentium/200 processor, 128 meg of RAM, and
a 2-4 gig disk drive, as long is it was only running one or two
server applications.
Linux was providing much better performance, at much lower cost,
on much smaller machines, and running many more applications
concurrently. Red Hat 4.2 was even named Product of the Year,
with NT 4.0 not even getting the honorable mention.
Microsoft knew that Linux was still growing too fast, so they
announced NT 5.0 shortly after they released Service Pack 3.
By September of 1997, Microsoft was promising NT 5.0 in early
1998. The problem was that Linux as a moving target. While
Microsoft kept adding bells and whistles like DCOM, COM+, and
ASPs, Linux just kept chugging along. What was worse was that
many of the third party developers who had previously helped
create the success of Microsoft were porting products like
WordPerfect to Linux. Many UNIX vendors even began porting
applications to Linux. Applix, who had been providing Office
Automation and Real-time Analysis software for Solaris, AIX,
and HP_UX for years was suddenly selling thousands of copies
of Applix to the Linux market at $100/copy.
Microsoft realized that it's desktop market was in jeopardy.
They again tried to lock out Linux by adopting the Universal
Serial Bus (USB) and DVD-Rom as key strategic features. They
even went all the way to the standards committees and told them
that unless every participant signed nondisclosure agreements,
that they wouldn't support USB. The DVD group wanted to prevent
Piracy, and Microsoft convinced the DVD group that it would be
better if they didn't try to patent the technology (which would
have led to disclosure to the UNIX community), instead they
should keep everything a trade secret.
Ironacally, this strategy also backfired. Linux users knew that
DVDs wouldn't work with the available versions of Linux, so they
specifically requested CD-ROMs instead. This raised a red-flag
with OEMs who suddenly realized that the Linux market was much
bigger than anyone had figured. It also gave an indication
of how fast that market was growing. Many of these Linux
users were still running Windows much of the time, but there
were a lot of people who were booting up Linux when they could.
Meanwhile, Microsoft's attempts to prove that NT was superior to
Linux were also becoming a problem. Microsoft began posting
benchmarks, press releases, and forcasts that were supposed to
show the superiority of Windows NT 4.0 to Linux. The Mindcraft
Benchmarks, the ROI and TCO estimates based on assumptions of
1 Sun 450 and one Novell NetWare machine against 3 NT machines
for 1000 users gave NT a better TCO by 30%. The errors in
the reports were being illuminated on the internet, and then
receiving widespread coverage in the mass media.
Microsoft had hoped to build a market for NT 5.0 by offering corporate
customers, MCSEs, and other key people free upgrades to NT 5.0 if
they purchased NT 4.0 right away. Microsoft was acutely aware of
the size of the Linux market, and was doing every thing they could
to make sure that the Linux user base didn't outgrow the NT user base.
At one point, Microsoft was even giving away copies of NT 4.0 to
members of MSDN.
Microsoft knew that Windows 2000 had to be a hit not only as a server,
but also on the desktop. They improved compatibility with Windows 95,
they improved reliability compared to Windows NT 4.0, and they
lowered the prices of Windows 2000, including Windows 95 users in
the qualifications for the "Upgrade Price".
As usual, Linux is still a moving target. Linux developers eventually
cracked the DeCSS code and Microsoft attempted to make it appear that
it was being used as a piracy tool. Linux developers have been
deciphering more and more of the USB codes, and Fire-Wire proponents
are throwing their support behind the Linux community. The Linux
community is still growing, and as more and more applications and
functionality are ported to Linux, more users are using Linux more
of the time. Many OEMs are now making sure that their hardware,
especially modems and video cards, are Linux compatible. Many
OEMs like IBM, Dell, and HP are now offering Linux compatible
machines. Compaq has been sitting on the fence. On one hand,
they have Jon "Maddog" Hall, True 64, and some very powerful UNIX
systems from DEC and Tandem. On the other hand, they don't want
to risk having their Microsoft license revoked again.
Even more interesting is that we now have companies like Cobalt
and VA Linux taking markets away from "Pure Microsoft" vendors,
and we have "Bare Bones" machines and used equipment being snapped
up by those who don't need Windows. Because Linux systems can
be used additively (a single workstation can tightly integrate
multiple application servers), many Linux users are setting up
multiple Linux systems.
Microsoft is also caught between a rock and a hard place on
the desktop market because the Judge has ruled that Microsoft
has a Monopoly in the desktop PC market. This is in fact
still true. The reason that many of these Linux users must
still run Windows much of the time is because Microsoft still
has control of corporate standards used for the exchange of
word processor documents, spreadsheets, presentation graphics,
charts and graphs, e-mail, and databases.
Microsoft now has a simple choice. They can begin to promote
open standards without extensions, such as XML/XSL, RTF, GIF,
JPEG, CGM, and other open standards, they can make their standards
public with sufficient detail to allow implementation using Linux
APIs and infrastructure, or they can take their case to the Supreme
Court.
At this moment, the server market is the one place where Microsoft
still has the right to compete without restriction. It is quite
likely that Microsoft is hoping that if it can gain control of
the corporate server market, and substantial control of the
Internet Server Market, that it can impose standards and protocols
that will still force end-users to stay with Windows instead of
switching to Linux.
> http://www.unix-vs-nt.org/kirch/
> "Why Windows NT Server 4.0 continues to exist in the enterprise
> would be a topic appropriate for an investigative report in the
> field of psychology or marketing, not an article on information
> technology. Technically, Windows NT Server 4.0 is no match for
> any UNIX operating system, not even the non-commercial BSDs or
> Linux."
>
> --
> Mr Rupert
>
>
--
Rex Ballard - Open Source Advocate, Internet
I/T Architect, MIS Director
http://www.open4success.com
Linux - 60 million satisfied users worldwide
and growing at over 1%/week!
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: [EMAIL PROTECTED] (Andy Newman)
Subject: Re: Nice link
Date: Sat, 1 Apr 2000 11:56:35 +1000
Reply-To: [EMAIL PROTECTED]
In article <ygcF4.1393$[EMAIL PROTECTED]>, Erik Funkenbusch wrote:
>Uhh.. did it BSOD or did it freeze? You have never been able to 3-finger
>from a blue screen.
The program is trying to read from the serial port and is waiting.
It shouldn't kill the OS - even Win9x - unless there is something
really screwy with the machine/configuration/whatever. It's an old
problem caused by the MS-DOS compatibility crap of making devices
appear in all directories. Programs have to check filenames to
ensure they don't use devices by mistake (yet another mis-feature of
Windows that leads to larger programs). Annoying, yes, but easy to
cope with but you have to remember to do so. BTW it caused all sorts
of trouble with the comp.unix.aux newsgroup on certain NNTP servers
running under Windows several years ago.
--
Chuck Berry lied about the promised land
------------------------------
From: Daniel O'Nolan <[EMAIL PROTECTED]>
Subject: Re: Top 10 reasons why Linux sux
Date: Fri, 31 Mar 2000 23:10:28 +0200
> Either way Linux doesn't provide a way for those AOL users to use AOL in
> Linux.
I gotta ask: Is this necessarily a bad thing?
-Dan O'Nolan
------------------------------
Date: Fri, 31 Mar 2000 21:08:37 -0500
From: Robert Morelli <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Crossposted-To: comp.os.ms-windows.nt.advocacy,comp.os.os2.advocacy
Subject: Windows 2000 has "issues"
How shall I say this delicately? Well, to borrow Microsoft's own words,
Windows 2000 has "issues." Now, let's not be judgemental of those less
fortunate. Every OS has emotional needs, and Windows 2000 is not the only
OS with problems "adjusting." DOS had "issues." Windows 3.1 had "issues."
Windows 98 and Windows CE have "issues."
One "issue" that Windows 2000 has just unrepressed, is a killer defect
that "poses a potentially devastating threat to ISPs, point-of-sale (POS)
systems and small- to midsize businesses." See,
http://www.crn.com/dailies/digest/breakingnews.asp?ArticleID=15279
It seems Windows 2000 has an "issue" with IP addresses that only allows it
to handle up to 51 of them. Remember how Windows 3.0 had an "issue" with
icons that only allowed it to put 26 in a folder? Remember how we
speculated that the programming wizzes at MS used the letters of the
alphabet to index the icons? You don't think they're using the weeks of
the year to index IP addresses, do you? Who knows? Maybe something
traumatic happened in that 52nd week, which would explain everything.
You know, I got an electronic rolodex once for $14.95 that could only
store 50 names and addresses, and the next year they came out with one that
could do 100. So, if you thought it was a cool idea to migrate from Solaris
to Windows 2000 and you're business depends on this, don't panic. They'll
probably have the 100 IP version out next year.
By the way, this isn't one of the 63,000 bugs Win2000 shipped with. Get this
clear; it's an "issue," not a bug. The mere fact that an undocumented
limitation causes your server to crash, doesn't make it a bug. In fact,
this "issue" was mistakenly reported to Microsoft as a bug by at least two
of its customers. Microsoft took 5 days to acknowledge the "issue," after
which an MS rep informed:
"Microsoft would not likely produce a hot fix for this, given that none of
our customers have reported the issue. If a customer does report this,
however, we will take it very seriously."
May I ask that we try to be compassionate? Yes, Microsoft is in denial.
And yes, they may lash out at those who threaten their defensive facade.
But try to imagine how much hurt and insecurity they feel. On the outside,
Windows 2000 pretends to be mighty and stable and scalable, but inside
it feels as infantile and unstable as NT. Remember, there is usually a stage
of denial that precedes the healing process.
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.advocacy) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Advocacy Digest
******************************
