[SLUG] Re: OT: M$ deals the final blow -- MS repeat of 10 years ago ...

From: Bryan J. Smith (b.j.smith@ieee.org)
Date: Thu Sep 09 2004 - 01:05:17 EDT


On Wed, 2004-09-08 at 22:13, Steve Szmidt wrote:
> You appear to really like MS the way you always seem to try to protect them.
> Did they not tell you - They ARE the evil overlords! Well, maybe not evil.
> But clearly don't have either your's or mine best interests at heart. It's
> all about getting more money in the bank at whatever the cost.

Ironically, this is true, but not because they are "evil."

As an original NT 3.1 beta tester at the largest site of the first
native NT 3.1 application, in 1994 when Gates through his weight behind
"Chicago" (MS-DOS 7 aka Windows 9x), Win32 was doomed. That was done
because 90% of the consumers were willing to expense a good platform for
compatibility. As such "Cairo" became impossible -- and the signs where
their as early as 1993, and 90% of the "technologies" _never_ came to
be.

This has now repeated itself in 2004 with Longhorn. .NET, an API that
would address all the weaknesses of today's "bastardized Win32"
(_nothing_ like original Win32), much like Win32 tried to do 10 years
ago. There were signs in early 2003 with the Alpha-quality developer
tools that Microsoft was going to pull back -- because the tools simply
were _no_where_ to support the OS that they were planning. Sure enough,
Longhorn is now just NT rehashed, and Blackcomb will _never_ feature 90%
promised.

But that doesn't make them "evil." They are a business, and if 90% of
consumers are willing to accept it, I can't blame them. In fact, the
_primary_ reason why Microsoft does this is because Microsoft's own
application developers don't want to change. They are _still_
supporting VB6, which forces the problem. Consumers are all too ready
to accept the problem.

Case-in-point: When the typical consumer goes to a superstore, they
have been programmed that anytime they upgrade Windows, or buy a new
computer, or want a new peripheral, they have to upgrade the other 2 in
the trifecta. That's just ludicrious to me (especially the engineer in
me), but it's what 90% of consumers believe.

Companies don't consider where their data will be in 2-3 years. They
only think about now. People don't care about security or reliability,
they want "ease of use." As someone who has used StarOffice since
versions 3/4 in the mid-'90s (back when it had HTML export and all sorts
of advanced features over MS Office 95/97), and can _still_ read my
documents verbatim today, I have _no_pity_ for people who complain.

Much less a lawsuit brought forth by a legislator from Utah, supported
by competitors that _only_ benefited competitors in the end (no
surprise). Who cares if IBM, Sun and others can "license" Microsoft IP
under terms not even the IEEE would accept as "fair and uniform."
That's the problem.

>From a TCO standpoint, the main problem with Microsoft solutions is
integration. The integration of inter-related BS just kills me.
Putting more and more into the kernel -- a problem that started in NT
3.51, balooned in NT 4.0 and came to the final "crack and hack" in NT
5.1 (XP/2003) -- plus all the Internet Explorer bloat, DLL requirements,
continued use of Int20-3Fh and Win16/Chicago APIs, etc... -- all because
Microsoft _failed_ to adopt Win32.

Again .NET and Longhorn is just another repeast.

All the meanwhile, with GLibC 2 (circa 1996) and GCC 3 (circa 1999+),
the GNU platform known as Linux is pretty damn binary compatible with
itself now. Add in the simplistic, but well-proven approach of
UNIX-like systems in "always multiuser," "everything's a file (and has
ownership/permissions)" and "networked for TCP/IP from day 1" and it
works pretty damn good.

Furthermore, unless you patch the kernel, you don't reboot. Everything
is piecemeal, complies with open interfaces -- so developers and
sysadmins alike can change just what they need to. That's why you can
take a service down, upgrade, bring it back up -- but easily change back
if something doesn't work.

That's why Microsoft won't ship a "pure" .NET OS this time around, like
it tried to do with Win32 10 years ago, but only ended up hacking the
heck of out it for "Chicago API" compatibility (defeating the purpose in
the end). If vendors have to re-write their apps with a major change in
binary compatibility, they might as well go Linux which has proven to be
_far_more_stable_ and _proven_ from an API standpoint.

Heck, even Miguel and Ximian have shown that .NET is better under the
well-established GNU framework than the bastardized Win32 API we have
today.

Which by 2006-2007, ironically, the most "advanced .NET experience"
won't be the WinFX technologies in the Avalon desktop or Indigo services
-- which are just basic .NET framework built upon and sandboxed over
hacked Win32. It will be in Mono and GNOME 3, which leverages a very
agreeable and proven GNU framework that _always_ enforces access
privileges, requires ownership and permissions, etc... on the smallest
detail.

As far as Avalon, when NT 4.0 adopted the "Chicago" GUI, Microsoft had
to hack the GDI into the kernel for performance. The same is happening,
with the related and reduced stability for NT 6 Longhorn. And as far as
sandboxing Indigo services atop of Win32, Microsoft has the same issue
with running Java atop of Win32. The underlying cracked'n hacked Win32
API is a Pandora box of bypasses and direct access.

> They don't just fix bugs. They introduce new ones every time.

Because they hack consumer compatibility into Win32. Why? Because even
their own application division and tool developers still use direct API
calls and memory access.

> Then they use mix up bug fixes with feature changes that leads to even bigger
> problems.

Which was _always_ the biggest complaint from NT developers in the
mid-'90s. With 1/10th the development team of "Chicago," thanx to
Gates' "vision" -- as I always say, "NT became DOS' bitch."

> No Robert, I think it's you who should stop protecting MS. I mean really, this
> is a Linux list with Linux users. Most of who are really fed up with the crap
> this multi multi billion dollar company happily throw upon us.

I don't think _anyone_ is "protecting" MS.
I merely explain why they do the things they do.
I blame consumers who are willing to put up with it.

> What do you expect to hear here? How great windoze is? You cannot expect to
> stand up for MS and not get shot at on a Linux list...

Community developed software built the Internet.
Community developed software built the world.

There were commercial entities involved.
There is plenty of room for even, what I call, "Commerceware."

Unfortunately, 90% of consumers are more than willing to pay for
"Hostageware" and "Superstore hardware." No wonder AOL, Microsoft and
many others own major percentages of them!

> Bah, that's nonsense! Why? Because you cannot "pass the buck" in security!
> The biggest problem is the unrepairable win32 API (programmer development
> tools). And the neverending ignorance of security issues that has led us into
> the quagmire.

A track back through the "evolution," or what I like to call "bend over
and take it NT" (from "Chicago"), is a good lesson in why Win32 sucks
today.

.NET was a great idea. But like many great APIs from Microsoft, they
don't even adopt it themselves. Why? Because their on developers don't
want to.

> And breaks everything while at it. No corporation with experienced admins
> allow autoupdate. You probably have never been in the position where you are
> responsible for the server farm to stay up, with the need to test changes
> done on a seperate LAN, so that you don't crash production servers. Something
> EVERY SINGLE SP EVER COMING OUT OF MS HAS DONE TO DATE.

XP SP2 is a marketing campaign to say "hey, we tried."

But even the new security chief of MS stated in February of 2004 that
_no_ version of Windows was _ever_ designed for the Internet. [ I'd
point to the PC_Support archives with the article, but they are down
right now. ]

But that's back when Microsoft was still putting up the "ruse" that
Longhorn was going to be .NET. I've known that was _impossible_ since
early 2003 -- just like back in the '90s with NT and "Cairo."

History repeats itself.

The bastard we now call Win32 is here to stay. Broken security model.
"Root privilege by default."

> So does leaving the computer off.

Firewalls are nothing. They provide _minimal_ protection.

> Come on Robert, do you really believe all that MS says? Big, money making, and
> successful has proven not to be the same as trustworthy.

The "Trustworthy Computing Initiative" is a marketing campaign.
Microsoft can do _nothing_ to change its own developers. For the first
6 months, security auditors were getting _locked_out_ of meetings. They
were bashed by application and tool developers. And Microsoft itself
can't do away with poorly written VB6 and its code.

The age-old "release date" argument is always used everytime.

I know. I was at a Fortune 20 company auditing bank code as a top-level
security analyst. You can prove to them how to the code is exploitable,
you can convince executives that it should not be put into production.
But they will argue to months not to change it.

I mean, why should they care? They won't be around to support it.
Companies are using contractors and outsourcing so much -- they piad
developer could care less.

Which brings me to my final point. As someone who has personally worked
with interviewed with Microsoft over the past 12 years, you get to see
how the company really operates.

They are an investment company -- not a software company -- that buys
software, outsources both development and IT, and uses _very_little_ of
their own products that they sell to enterprises every day. It's rather
sad.

So the way they operate is 100% understandable. Evil? Hardly. They
have a very naive consumer base.

-- 
Bryan J. Smith                                  b.j.smith@ieee.org 
------------------------------------------------------------------ 
"Communities don't have rights. Only individuals in the community
 have rights. ... That idea of community rights is firmly rooted
 in the 'Communist Manifesto.'" -- Michael Badnarik

----------------------------------------------------------------------- This list is provided as an unmoderated internet service by Networked Knowledge Systems (NKS). Views and opinions expressed in messages posted are those of the author and do not necessarily reflect the official policy or position of NKS or any of its employees.



This archive was generated by hypermail 2.1.3 : Fri Aug 01 2014 - 17:01:20 EDT