Linux-Advocacy Digest #866, Volume #25 Wed, 29 Mar 00 00:13:05 EST
Contents:
Re: What should be the outcome of Microsoft antitrust suit. (R.E.Ballard ( Rex
Ballard ))
Re: Windows 2000: nothing worse (Jim Richardson)
Re: Why did we even need NT in the first place?
Re: Multithreading ("Marc E. Christensen")
----------------------------------------------------------------------------
From: R.E.Ballard ( Rex Ballard ) <[EMAIL PROTECTED]>
Subject: Re: What should be the outcome of Microsoft antitrust suit.
Date: Wed, 29 Mar 2000 04:45:41 GMT
In article <2mcE4.569$[EMAIL PROTECTED]>,
"Erik Funkenbusch" <[EMAIL PROTECTED]> wrote:
> R.E.Ballard ( Rex Ballard ) <[EMAIL PROTECTED]> wrote in message
> news:8bqurt$hal$[EMAIL PROTECTED]...
> > Bad idea. Most customers don't have the
> > technical expertise to install
> > off-the-shelf versions of any operating system,
> > including windows, without the special OEM files
> > provided by the manufacturer.
> I'm glad you recognize this, since it
> conflicts with parts of your later
> statements.
I've generally asserted that there are a substantial number of
people who can (perhaps 35-40%) install Linux with the aid
of appropriate supports (prompt files or configuration settings
provided by the product makers). I had no trouble installing
Corel, Mandrake, and SuSE on my ThinkPad. I don't have the sound
cards up, or the winmodem, but I found a cheap real-modem and
the sound drivers are available but it's not that high a priority.
My son had no trouble installing Mandrake on his home PC, once we
installed a Linux compatible card. I paid $15 for the card, and
it took 15 minutes to switch.
> > So getting to the bottom line: What should be included in the
> > settlement.
> >
> > Business Practices:
> >
> > Pricing schedule:
> > Microsoft should be allowed a spead of 20% on no more than 50%
> > of the previous year's sales. If the manufacturer orders more
> > than 50% of the previous year's sales, they will get no further
> > discount. The Federal Trade Commission would be given the right
> > to monitor all contracts to make sure than the spread does not
> > exceed 20%.
>
> So now you are penalizing an OEM for
> increasing their sales by more than 50%.
No. The OEM still gets the 20% discount because he's a big
volume customer. He just doesn't get MORE than 20% off.
The problem with the 1994 agreement was that Microsoft reserved
the right to set it's own pricing schedules and then used this
practice to circumvent the clause which forbade "per processor
licenses". Since Microsoft couldn't force the OEMs to put Windows
on each machine, they made it cheaper for the OEM to purchase 120% of
what he could possibly need than to purchase 80% and order more at
the same discount. Of course, if the OEM overestimated his need,
part of the process for getting credit was to show that they had
installed Windows on every machine.
Microsoft also punished OEMs who didn't accept all of Microsoft's
terms and who didn't install Microsoft tie-in software instead of
a competitors. IBM was charged as much as 9 times more because
it refused to stop installing Lotus Smart Suite.
Finally, Microsoft threatened to revoke licensing agreements
on an "all or nothing basis" because Compaq refused to accept
the conditions that prevented the installation of Netscape as the
primary browser on the desktop. Keep in mind that Compaq hadn't
uninstalled IE, they had just taken it off the desktop and put
Netscape in it's place.
> If I buy 1 copy last year, and this year I want 100,
> I have to pay full price for 98 of them?
No, you'd get the 20% discount, because you purchased more than
50%. You just wouldn't get an 80% discount for buying 120 of
them (which would then mean you pay less for 120 than for 80).
> > Tie-ins:
> > Microsoft will not be able to offer discounts in exchange for
> > non-monetary compensation such as the inclusion or exclusion of
> > competitor productes. Each component much be independently
> > managed and marketed.
>
> So, Microsoft could not offer discounts to OEM's that, say,
> follow the latest PCxx standard?
Nope! Especially that! Microsoft want to dictate hardware
standards and shouldn't have the right to do this. This is
especially true of protected standards like USB and DVD-CSS.
If it's in the OEMs best interest to produce PCxx machines, then
Microsoft doesn't need to offer a deeper discount. If instead,
the OEM could make a better or cheaper machine by using Fire-Wire
and DVD drives that can be used with Linux (CSS encription on-chip),
then they shouldn't fear a reprisal. This is how Microsoft
maintains it's monopoly. Pull out this prop and the market becomes
more competitive. In most cases, Microsoft rewards OEMs who
use Microsoft-exclusive technology and punishes OEMs who use
Linux-inclusive (competitor inclusive) technology.
> They couldn't offer discounts to those that get their
> software certified (and most larger OEM's create
> their own software for a number of things)
> instead of just producing buggy garbage?
Nope! If an OEM wants a really good driver, he can hire a
Microsoft consultant at $280 an hour to help out. If the
OEM can't get a working driver in a few months, then the
problem is with Microsoft, not the OEM. Perhapse the OEM
would rather market Linux on that hardware.
Again, the goal is to create a competitive market in which
Microsoft no longer has the monopoly control. When Microsoft
is the only game in town, you must accept Microsoft's terms.
When the OEM has a choice of an operating system published
by an uncooperative provider (Microsoft) and an operating
system published with complete source code and debugging tools,
Microsoft will have no choice but to be more cooperative and
competitive.
> > Innovations:
> > Microsoft may not provide any incentives for the inclusion of
> > products for which there are other competitors. For example,
> > Microsoft may offer Internet Explorer as a complimentary product,
> > but may not make the inclusion of it, or it's placement on the
> > desktop a requirement for aquiring the operating system.
>
> Well, guess MS would have to get rid of the calculator,
> wordpad, notepad, edit, more,
> the CD command (since commercial versions of these exist).
First of all, I said that Microsoft could not force the inclusion
of these things as a condition of the operating system installation.
In most cases, the OEM would simply include Microsoft's tools. But
they could choose instead to install an updated version of the
Borland (now Inprise) versions. Microsoft would not be able to
demand that only the Microsoft versions could be installed, nor
could they demand that their versions be installed or left installed.
It's possible that an OEM might want to install the qt versions,
which are compatible with KDE implementations.
> It couldn't even ship with a command line shell
> since that would put hardship on 4Dos.
Microsoft could provide a command line shell, but they couldn't
prohibit an OEM from installing a fully functional Korn shell,
or PERL, or even EMACS. If the OEM wanted to install the entire
MKS toolkit, Microsoft wouldn't be allowed to prevent it.
Furthermore, Microsoft might need the MS-DOS command.com shell
for it's internal scripts, but that doesn't mean that the OEM
can't move the shortcut to a more or less prominent part of
the desktop (in the Accessories folder for example). The
OEM might also want to put IE5 in the accessories folder as well.
> > Predatory coding:
> > Microsoft may not create code that would damage any existing
> > subsystem whether provided by Microsoft or any competitor. This
> > includes any modifications of the Master Boot Record, the
> > partitioning, or the release of upgrades that damage competitor
> > products. When there are disputes over predatory code, Microsoft
> > will submit the code in question. The arbiter could be the FTC.
>
> So, how would those cluless newbies you
> mentioned earlier figure out how to
> modify their own MBR?
If a Linux system was already installed, Microsoft would have
to either honor the Linux partition, or offer the user the option
of either leaving the MBR as is (meaning that Microsoft would have
to add instructions for setting up Windows on Klilo and LILO), letting
the user use Microsoft boot manager that could boot both systems,
or installing Windows over Linux. The key is that it would always
be the user's choice. Furthermore, if Microsoft damaged a user's
system without the user's direct interaction and without warning
the user of the consequences of taking such an action, it would be
the same as a hacker sending out the explore-zip virus. The
code in question was posing as benovolent and useful code, but
willfully damaged an existing system.
Under the computer tresspassing act, programs that format hard drives,
remove user files, and mess around with boot tracks are normally
treated as criminal activities. A Windows 98 installation procedure
that damages the code already installed on a system is merely a very
expensive virus.
The distributors of other operating systems, in a competitive
environment have had no trouble adhering to the law. They make
every effort to preserve the integrity of the previously installed
software. This is true of BEOS, Linux, and BSD. I haven't tried
installing Solaris for Intel on Windows 9x or Windows NT system.
> > The Findings of Fact Stand:
> > All portions of Judge Jackson's findings of fact are accepted
> > by Microsoft. Microsoft will be given amnesty for previous
> > conduct, but will be treated as a monopoly, and all testimony
> > and findings can be used in the event of future misconduct.
>
> There are things in the findings of fact
> that are proveably wrong.
Then Microsoft should take it's chances with the ruling. If
Microsoft wants to "cop a plea", they need to accept the evidence
presented and accept the judges Finding of Fact. At this point,
Microsoft is seeking to avoid a verdict which may lead to both
civil and criminal charges. It's like the man accused of murder
who could possibly refute some of the evidence, but accepts a
lesser charge of manslaughter rather than risk execution.
> For instance, this would essentially make
> it law that Apple could not compete
> with Microsoft for the same market.
No. It merely define the market in which Microsoft had a monopoly,
and explained why Apple was not currently a legitimate competitor
in that monopoly. Microsoft presented it's case that it was a bit
player in a larger market, but the Judge defined the nature of
the monopoly. He ruled that Microsoft controlled the Intel PC
market, both through the application base and through the cost
of replacing the hardware. He noted that while Apple also made
computers, that the cost of replacing the hardware so that users
could run Apple's operating system too prohibitive - thus another
barrier to entry.
If Apple decided to make an Intel version of OS/X, and then captured
30% of the INTEL PC market with this operating system, then Microsoft
would no longer be a monopoly and would no longer be subject to the
Sherman act and Clayton act.
> The FoF is also misleading in many places.
The Judge makes the final determination based on the evidence
presented in the courtroom. Microsoft had it's chance to present
it's side of the case, and often made an embarrassing spectacle of
itself. Fortunately for Microsoft, most of the country was too
busy following the Monica Lewinsky scandal first leaked by the MSNBC
web site, covered nearly 8 hours per day during prime-time by MSNBC,
and nearly 4 hours per day by CNBC.
Microsoft used it's affiliation with NBC to make SURE that the
Microsoft trial received as little coverage as possible.
> The testimony about OS/2 completely ignores
> the fact that IBM was a competitor to other OEM's,
Irrelevant. Microsoft attempted to prevent IBM from putting
OS/2 an it's OWN machines. IBM might have become a viable
competitor, but Microsoft stonewalled IBM up to 15 minutes before
the "Big event" unveiling. Even this was only done so that Microsoft
could say list the IBM logo on it's list of Windows supporters.
> and that most OEM's wouldn't put OS/2 on their
> systems if they were paid to do so,
> regardless of any MS contracts.
That wasn't the issue. The issue was that IBM was prevented
from marketing it's own product. Microsoft admitted to issuing
an ultimatum that IBM either stop selling OS/2 or they wouldn't
get Windows 95. When that didn't work, Microsoft admitted that
they demanded a license audit and demanded a $30 million payment
along with an agreement that IBM stop promoting OS/2 before any
further negotiations could take place.
It's entirely possible that IBM would eventually have dropped
OS/2 voluntarily, but instead IBM wasn't even allowed to tell
it's customers what Microsoft had done. They just simply stopped
advertising OS/2, stopped shipping OS/2 systems to retailers, and
stopped direct-selling OS/2 to customers except when requested
by the customer. IBM literally killed the demand for a product
which it had purchased at a very high price from Microsoft.
> > Nondisclosure Agreements:
> > When a competitive standard is provided in a publicly available
> > formate, Microsoft may not use nondisclosure agreements or
> > incentivess for creating alternative proprietary standards.
> > All future communications protocols will be regulated by the
> > Federal Communications Commission. The use of proprietary
> > hardware interfaces will be regulated by the Federal Trade
> > Commission. This will only apply to protocols and standards
> > that are not available under the same terms as competitive
> > standards, and will only apply to standards bodies in which
> > Microsoft is a contributing member. Microsoft may not use
> > nondisclosure agreements to prevent the disclosure of
> > information that could otherwise be copyrighted or patented.
>
> I assume you're talking about Java here.
Java, CORBA, HTML, XML, TCP/IP, DNS, ARP, LDAP, X.509, SCSI,
FireWire (IEEE), NFS, HTTP, Postscript, TeX, SGML, CGM, GIF,
JPEG, MPEG, ext2, tar, lf delimited text, ... all components
of POSIX level 3, UNIX 95... all components of Linux...
> Java is not, nor is it likely to ever be,
> a standard other than a defacto one.
BINGO!!!
Furthermore, the defacto seems to be JDK 1.1.8 rather than JDK 1.2
> Even then, Sun does a great
> job of violating it by itself in
> it's various version changes.
Yes. This is why JDK 1.1.8 has been widely adopted and JDK 1.2
or 2.0 are still in the background. Even EJB enjoys limited support
and depends on 1.1.8 backward compatibility.
> > Deceptive Marketing:
> > Microsoft will not be allowed to use nondisclosure agreements
> > to suppress reports which are accurate but unfavorable to
> > Microsoft. This includes, but is not limited to, benchmark
> > results, bug reports, feature comparisons, or comparisons to
> > competing technology.
>
> With the exception of beta software
> (which NDA's should apply to, since the
> results can change),
But the terms of the beta agreement do not terminate with
the general release software.
> the NDA's you're referring to are signed in order to
> get free versions of the code for testing.
Actually, free or discounted versions. For example, you
are bound by such terms when you get the combination of
MSDN Enterprise and a "Strategic alliance agreement".
Microsoft refutes the credibility of anyone who does not have
a strategic alliance agreement. They can't have it both ways.
They either have to allow publication of whatever facts may
be relevant, or they have to accpte the results conducted
by independent testers - which may include Linux zealots.
> If they want to do non-biased benchmarks, they could buy it.
It's very hard to find anyone whose company hasn't signed a
strategic agreement with Microsoft that contains this restricted
disclosure clause.
For nearly every single previous employer, I was able to purchase
my own software at standard CompUSA prices, run my own benchmarks,
and even publish my findings within the company, but I was not
allowed to disclose these result to outside sources without
Microsoft's permission. At each employer, I was also able to
find other independent benchmarks which were unfavorable to
Microsoft, but were classified as company confidential. I can't
even talk about them now.
Finally, even though I was able to make back-ups of some systems,
it would be a violation of my previous employment agreements to
publish certain Microsoft related information on my web site.
> > Conditions for lifting the restrictions:
> > The restrictions listed above would be lifted once competitors
> > have penetrated at least 50% of the market. This includes both
> > coexistant (dual-boot) systems and competitor only systems.
> > This 50% will only include those platforms on which Microsoft
> > Operating Systems are licenced and the regulations would be
> > lifted on a market-segment by market-segment basis. Thus,
> > if Microsoft controls less than 50% of the hand-held market,
> > there would be no restrictions on that market, but the existing
> > restrictions would still exist in the Intel PC market.
>
> Microsoft is not being tried for any
> other market than PC operating systems.
> Therefore, any agreement should only apply to that.
Sounds fair. The exception being that Microsoft might attempt
to use it's current monopoly of Personal Computers to extend
it's monopoly of hand-held or appliance systems.
The key is that Microsoft can't count palm-pilots and say that
palm has 50% of the PC market. They can't count appliance
machines such a the QUBE or Cisco Routers and say that Cisco
has 50% of the market.
The key factor is that Microsoft can't unilaterally redefine
the market and use that redefinition a case for lifting the
restrictions.
If the DOJ/FTC decides to include appliance workstations and
NCs or LCs as part of the overall market definition, that
is something that both sides would have to mutually agree to.
This might even mean that Microsoft wouldn't be regulated in
the Internet Server market (since they only have 25% of that
market). I'd leave that one up as a bargaining chip.
> > The 50% rule also applies to all competitive products installed
> > on an after-market basis.
>
> Huh?
If somebody comes up with a franchise that installs Linux on PCs
at a local Kiosk, and that suddenly drives the Linux market so
high that Linux is sitting on over 50% of the PCs being sold,
then Microsoft would no longer have a monopoly. They would
have a viable competitor (or two or three) and restrictions
would no longer be neccessary.
I really don't think that we need to regulate Microsoft forever.
Eventually, when we start seeing retail PCs displayed on the retail
shelves that run Linux as well as, or instead of Microsoft Windows,
consumers will be making their own choices. Microsoft might still
be the dominant player - with 4 linux distributors, BEOS, and 3
BSD distributors all competing for pieces of the market on a coresident
or exclusive basis.
Once this happens, the nature market forces will encourage the adoption
of open standards and Open Source (or multiplatform) solutions that are
not entirely dependent on the whims of Microsoft. At this point,
if Microsoft tried to say "you need us, you can't survive without us,
and you must stop selling anything but our products", the OEMs would
laugh, and ask them whether or not they wanted 30 or 40% of the market.
> > Penalties and Enforcement:
> >
> > If Microsoft violates the initial court decree, they would
> > be in contempt of court. Furthermore, executives ordering
> > actions which violate this agreement could be jailed for
> > contempt, fined based on a percentage of personal assets, or
> > prosecuted under criminal racketeering laws.
>
> That would be unconstitutional,
We jail noncustodial parents for failing to comply with the terms
of a voluntary civil settlement, often for up to 5 years without
even a trial. If the executives know that they could be jailed
for contempt, they might think very carefully before even thinking
about trying to tap-dance around the settlement, the way they did
in 1996.
> given that the Sherman Act is a civil,
> rather than criminal code.
Slapping Microsoft with a $1 million/day penalty isn't dealing
with the right problem. It hurts the investors, but Bill Gates,
Steve Ballmer, and Paul Allen own over 60% of the company. They
can't be fired, they can't be voted out by a board of directors,
and they can't be forced to resign.
My guess is that if Bill Gates had to spend a few nights
in a federal jail, he wouldn't even think about issuing a
memo demanding that OEMs ignore the terms of the settlement.
It's very important that the settlement be made public as well.
The OEMs, the Consumers, and the Competitors must know the terms
and be willing to "blow the whistle" if they even think that
these terms are being violated.
> > Microsoft would be fully liable for all damages to competitors
> > and for damages to the consumer market as a whole for any
> > violations. These would be subject to binding arbitration by
> > an appellate court and could only be appealed to the supreme
> > court.
>
> And tie up the legal system for decades
> while companies try to make fast
> bucks by claiming violations.
Yup! If there is a case, and Microsoft doesn't want to settle,
they can be hauled through the courts over and over until they
learn to behave.
Remember though, this restriction only stays in effect until
Microsoft's market share drops below 50%. After that, the
Judge and the DOJ would declare that Microsoft is no longer a
monopoly, and no longer subject to the Sherman act.
If Microsoft looses the market, then comes back with something
that will "blow the socks off of everybody", and the OEMs, competitors,
and consumers willingly put themselves back into being at the mercy
of the new Microsoft Monopoly, it was their choice. But the DOJ
would have a much harder time proving that the public was damaged
by that choice.
>
--
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] (Jim Richardson)
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Windows 2000: nothing worse
Reply-To: [EMAIL PROTECTED]
Date: Wed, 29 Mar 2000 05:02:19 GMT
On 28 Mar 2000 05:04:46 GMT,
Stephen S. Edwards II, in the persona of <[EMAIL PROTECTED]>,
brought forth the following words...:
>Donovan Rebbechi <[EMAIL PROTECTED]> writes:
>
>: On 27 Mar 2000 23:20:36 GMT, Stephen S. Edwards II wrote:
>
>: >Yet, the root account still exists on the system. You seem to have
>: >overlooked that little fact. The very existence of root is a problem.
>
>: Why is it a problem that it exists if it is not used ?
>
>I'm thinking in terms of malicious intent here. Disgruntled employees,
>etc. It has, and does happen often. I think this was the context of
>Erik's argument as well, but I shouldn't speak for him in that fashion.
>
>: >Erik isn't saying that root is too much power to handle. He's saying that
>: >it doesn't make sense that some sysadmin, drunk on power, can just
>: >carelessly delete data files of other users at random, if he/she so
>: >wishes.
>
>: If the systems administrators are "drunk on power", the users are pretty
>: screwed regardless. And as it's already been pointed out, you have the
>: same kind of issues with the Adminstrator account on NT ( take ownership )
>: The point is that *someone* has to have control over the system.
>
>Right. But there is also the point of making a mistake, which humans are
>known to do from time to time.
Which (as has been pointed out before) is a good reason to not run as
root.
>
>: Sure, you can abuse "rm". You can also abuse a hammer and bash the case
>: with it. The point being ?
>
>The point being, that with root and rm, you can wipe out an entire user
>database in seconds. With WindowsNT, you have to take ownership of each
>and every file and folder. This has to be done manually, and cannot be
>done by mistake, or on the fly.
So if Joe Dipstik runs some .exe he got in email, and it starts creating a
bunch of files in his home dir, or any other public space, you have to
change ownership of each and every file to kill them?
--
Jim Richardson
Anarchist, pagan and proud of it
WWW.eskimo.com/~warlock
Linux, because life's too short for a buggy OS.
------------------------------
From: [EMAIL PROTECTED] ()
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Why did we even need NT in the first place?
Date: 29 Mar 2000 05:03:55 GMT
In article <[EMAIL PROTECTED]>, Andrew wrote:
>Perhaps you didn't look at the software hard enough. If you installed Terminal
>Services you can remotely administer the machine with the full GUI. Yes, I know
>MS didn't package a Linux or Solaris client, but from what I understand, the
>Citrix ICA client is fully compatable with it (they use the same protocol). I've
>never done it with the Citrix client myself, so YMMV. As for W2K Professional,
>you can remotely administer NT servers decently enough for many duties with the
>Computer Management MMC plugin. These are just two built-in examples. NT/W2K has
>a way to go before it's perfect in this respect out of the box, but it's goten a
>HELL of a lot better with this release.
However the Terminal Services Package is not part of the standard W2K
release is it ?
>
>You knew the latter was not true and you obviously did not have enough time to
>test the former, so why did you even bother?
It is true for me if I can not run the software on Intel, Alpha, Sparc, Risc
I have to see it as a toy...because in my environment that is the hardware I
use and I can not spend time supporting a OS-type that is not working on
all those platforms. Look the day Windows scales from a Intel-386 to a
Sun E10000 I will take it seriously... Until that day I will stick to Unix.
You see ones you have your software developed you can scale it on Unix by a
simple re-compile..If I would choose NT I would have to rewrite most
applications thus wasting money. I'm one of those poeple who doesn't like
wasting money and thus also energy having to rewrite a whole application
because for example Intel systems don't perform well on IO. For this I can
state that a Sun Sparc 2 with 64 MB Ram outperforms a SMP Intel Box on Raw
IO Performance. Also my alpha & Risc system's perform better then Intel on IO
based stuff. You see I'm steadely disposing of my Intel based hardware and
replacing it with Real Hardware. The only system I will be keeping that is
Intel based is a SCO workstation for the rest I don't want Intel systems
anymore.
Michael
---
Michael C. Vergallen A.k.A. Mad Mike,
Sportstraat 28 http://www.double-barrel.be/mvergall/
B 9000 Gent ftp://ftp.double-barrel.be/pub/linux/
Belgium tel : 32-9-2227764 Fax : 32-9-2224976
------------------------------
From: "Marc E. Christensen" <[EMAIL PROTECTED]>
Subject: Re: Multithreading
Date: Tue, 28 Mar 2000 22:05:41 -0700
Bob Hauck wrote:
>
> On Tue, 28 Mar 2000 20:46:12 GMT, Erna Odelfsan <[EMAIL PROTECTED]> wrote:
>
> >does Linux still have to use green threads for Java or are there native
> >threads ?
>
> The IBM JDK uses native threads. I believe the latest Blackdown port does
> too, but I haven't tried it yet.
I also know that the IBM port uses native treads. Currently, it's also
the faster of the two ports. However, I expect that to change in the
future.
--
Marc C.
An Official LDP Mirror:
http://www.mecworks.com
------------------------------
** 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
******************************
