Linux-Advocacy Digest #939, Volume #27 Tue, 25 Jul 00 04:13:06 EDT
Contents:
Re: MS advert says Win98 13 times less reliable than W2k (R.E.Ballard ( Rex Ballard
))
Re: Samba vs NT, which gives best PCs / Server Performance? ("Stuart Fox")
----------------------------------------------------------------------------
From: R.E.Ballard ( Rex Ballard ) <[EMAIL PROTECTED]>
Subject: Re: MS advert says Win98 13 times less reliable than W2k
Date: Tue, 25 Jul 2000 07:01:50 GMT
In article <mrSd5.3443$[EMAIL PROTECTED]>,
"Spud" <[EMAIL PROTECTED]> wrote:
> [snips]
>
> "R.E.Ballard ( Rex Ballard )" <[EMAIL PROTECTED]> wrote in message
> news:8l7ei4$vte$[EMAIL PROTECTED]...
[part 2 - due to limits on DejaNews]
> Yeesh... $5,000 in labour for *what*?
Full backup of all user software (done by user) 10-30 hours
(identification of private files, back-up over 100/T network
in large corporation)
Removal of old hardware to desktop location 1 hour.
(latency between remover available and user willing to let go of
machine - can be up to 5 hours with executives).
Installation of new hardware at desktop location. 1 hour
(latency issues again).
Standard software configuration - 1 hour (waiting for
uploads from the network Installer + user).
Custom software configuration - 1 hour (software pre-authorised for
department).
Personalization (security settings, permissions, IP configuration,
multiple reboots, ...) 1-3 hours.
User Data recovery - 1-3 hours.
User Application Configuration - 4-7 hours.
User Data reconfiguration - 4-7 hours.
User retraining - 1-4 hours.
This gives you totals of 24 staff hours (best case)
to 59 hours (Mr Murphy working overtime). A pretty reliable
"rule of thumb" is to assume 40 hours of time lost (any combination
of installer and user) for each installation. This doesn't factor in
things like "playing with the toys", "grousing about the new system",
discovering your payroll and review information was read by an employee
while on the shared drive, and discovering that the upgrade made some
of your files unhealthy.
> Setting up new systems?
> Pathetic. If it takes you more than an hour to roll out a new
> machine, you're doing something wrong.
When you're budgeting for 1,000 or more users, assuming that the
1 hour taken to pop up the cover on the best machines in the house,
add some RAM, and install an upgrade onto a minimally loaded machine
running only the most kosher applications is a good way to give your
CEO and CIO a very unpleasant shock. Most CIOs don't drill into the
details of your estimates until it's too late. Most managers are
looking for the lowest number they can get, and most CEOs are
measuring the impact in lost sales, missed opportunities, late
projects, and litigation. A good installer figures he can
cuncurrently load 3-5 machines at a time, but he also realises
that users are very personally invested in the contents of their
Personal Computer (many users are like protective mothers turning
their child over to the doctor for high risk surgery.
> If you can't roll out a dozen
> of them at a time, you're doing something wrong.
In a "boiler room" where you have 30 machines sitting side by side
on a high speed lan and all you are doing is installing brand new
software on brand new machines to be assigned to new-hires, you're
correct. Even with all the reboots (thank GOODNESS Win2K doesn't need
to be rebooted every time you wipe your nose).
No doubt' Win2K is a bit cheaper than WinNT, but it's important
to realize what the real costs are. It's what happens BEFORE
you deliver the machine to the user (the user may plan your arrival
for several WEEKS), and what happens AFTER your installation (the
user will be recovering from the damage for several weeks as well).
> And if you *need* to
> roll out a dozen at a time just to upgrade, you've done something
> wrong in your initial purchasing... so don't blame someone else for
> incompetent business decisions on your part.
The *Newest* machines don't need replacement, but the owners of the
Newest machines are the ones who need the most horsepower. You
normally have to rotate the machine from the power-users to the
managers, from the managers to the secretaries, and from the
secretaries to the Mainframe Programmers :-).
And the planning that goes into upgrading a call-center is
spectacular. This is one of those jobs that takes 3-5 people
a few months to coordinate, then, usually over a week-end, you
have pandamonium as an army of workmen literally tear-down and
rebuild the entire call center in less than 68 hours (from 8PM friday
night to 5:00 A.M. monday morning). During the replacement, larger
screens are installed, better keyboards are installed, new
telecommunications equipment is installed (phones, headsets...).
The only thing that is more of a "Ballet of Chaos" is the set of
a touring company of a Broadway show.
> > with new systems, and recover back-up data into the existing
> systems.
>
> Oh, BTW, backing up and restoring existing systems isn't an issue.
It depends on a combination of corporate policy, the nature of the
PC being used, and the nature of the machine.
> You *are* aware of how to manage large-scale networks, right?
How big do you want to talk? Yes, I'm aware of the common practice
of sticking shared directories out there which users can use for
storage of non-confidential, non-private, non-secured, non-strategic
information which won't create huge liabilities if someone in the
administrators group doesn't "accidentally" pop them open and divulge
the contents. Usually the guy you were about to fire anyway - but
he's just collecting a little collateral, a few thousand passwords
he can use from a dial-in port, the stuff that you really didn't think
he was bright enough to do. He's the guy who will e-mail the payroll
figures to the company-wide distribution lists the week after you
thought you had locked him out. He also sent that letter to the CEO's
girlfriend to the CEO's wife (and the entire board), and if there
was any other little secrets, the'll get dumped, just a few at a time
until he either gets paid off, or bumped off.
> Hint: it's a lot cheaper and easier to drop a high-speed LAN
> in with a redundant server automatially backing up every night,
> than to fart around with per-PC data storage.
Hint: user data, corporate data, and software are so interwoven on most
Windows installations that it's nearly impossible to sort out which
items need to be saved and what's going to get clobbered by reinstall.
Most people have been burned badly enough, enough times, to carefully
sort through nearly every directory for personalized information that
has been stuffed into obscure directories all over the hard drive.
The more astute users put the user data in a "home" or "user"
directory (just like with UNIX), and keep copies of things like
bookmarks, shortcuts, cookies, configuration files, and such
in that directory. Still, just before the replacement, they will
spend several hours culling through the directory tree to make
sure they didn't miss anything, and several hours after the recovery
trying to recover things they did miss.
> If your users are storing things
> locally to their PCs, make it company policy that it is only *their*
> data, their personal data; company data gets stored on the network,
> where it's safe from drive crashes, gets backed up, and doesn't *need*
> to be restored on a per-PC basis.
UNIX and Linux has a pretty simple solution. They give each user
their own private directory, and everything goes somewhere on that
directory tree. Furthermore, the configuration files for each
application are stored in text files that can be backed up and
recovered if they are erased by a subsequent installation or upgrade.
Windows gives you the monolithic registry, which is nearly impossible to
manually edit, even harder to back up sequentially, and makes almost no
provision for manually selectable revision control. User configuration
information is stored in subtrees of each application, and all of the
software is stored on the user's local drive.
When you have 10 gig of applications, private data, user preferences,
and registry setting all woven together like a persian rug, back-up
and recovery get "interesting". As in "may you have an interesting
life" (ancient chinese curse).
> > If you want to follow the "rush", you'll pay the highest possible
> > price (around $10K) - and then you'll have to cut bonuses, raises,
> > and incentives by an average of $10,000 per employee. The Microsoft
> > lovers will be thrilled to get this new hardware (at everyone else's
> > expense), and most of the rest of the company will just figure it's
> > one more piece of software to learn.
>
> Don't know where you get any of this. It all sounds like extremely
> incompetent management and roll-out policies; you should be discussing
> with management how to get a competent network administrator in, not
> complaining about Microsoft. From the descriptions you're offering, a
> decent netadmin would save you at least 5 million a year... which I
> think will more than pay for all of your software licenses, with
> enough left over for a company party.
Which one. McGraw-Hill, Merril Lynch, Morgan Stanley, Dow Jones,
Microsoft, US-West, Citigroup, et al. By the way, this number also
happened to be the consensus given by Garner group for doing the
Windows NT upgrade (from Windows 95). Their estimate, based on
surveys of 300 of the Fortune 500 was $10,000. Because people have
been planning for NT 5.0 since 1997 (right after NT 4.0 was released),
the costs should be much lower, about half.
> > > "Windows-only hardware"? How do you manage that? As far as I can
> > > see, if you could write a Linux driver for a given piece of
> hardware,
> > > you could use it under Linux. If hardware vendors aren't
> providing
> > > Linux drivers, there may be a reason for it... perhaps a
> perception -
> > > correct or not - that Linux is but a bit player, not worth the
> effort.
> >
> > You create a proprietary protocol, like DVD-CSS,
>
> Ah, so you create Windows protocols, not hardware. Okay... then the
> vendors have the choice of adopting that protocol or rejecting it.
No they don't. If they want to adopt it, they have to accept all
of the terms Microsfot dictates, including some terms that may require
them to commit criminal acts (obstruction of justice, restraint of
trade, publication of false or misleading information, exclusion of
competitors -- to Microsoft...) and the IHV (independent hardware
vendor) must agree to all of it unconditionally and without public
objection.
The alternative is to not only be excluded from the new emerging
marketplace (created by Microsoft by virtue of it's ability to
put anything it wants into windows and then force every OEM to
swallow every byte of code without modification), but also to
risk having driver software written to support your hardware be
either excluded from the CD-ROM distribution, released with the CD
in such a buggy state as to render it useless (or worse - it causes
damage to the system) and then have Microsoft making public
statements that your company isn't really supported by Microsoft
- since you didn't sign the "Crazy Papers" (the papers that declare you
certifiably insane to your investors and employees, but prevent anybody
from talking about it).
> It's been done before - look at IBM and MCA, when the rest of the
> world told them to take a flying leap. (Okay, MCA isn't a protocol,
> but the principle applies.)
Yes it does. And when IBM tried to tie OS/2 and MCA together,
rather than fully supporting a public standard, and refused to
publish documentation for future pin-out definitions, the industry
balked completely. They opted for Windows 3.0 (buggy and crash-prone
as it was), and provided support for BOTH EISA and VLB (more expensive
but cheaper than the risk of being dead-ended by IBM).
When Akers and the "insider board" still didn't get it, and the value
of IBM stock dropped 60% in less than 3 months, it was a wake-up call.
IBM brought in an outsider (Lou Gerstner) who had been an unhappy
IBM customer, and who literally destroyed and recreated the entire
IBM culture. The NDH syndrome had to go, the reps were assigned
companies and were told to give the customer whatever the CUSTOMER
wanted, even if it was a competitor's product. Almost instantly,
IBM found themselves developing products for SunOS and Solaris
even before AIX or MVS, and they found themselves supporting Windows
even before OS/2.
When customers complained about the time required
to recover OS/2 2.0 IBM provided fast recovery tools
in Warp 3.0, and they fixed the "Trap 0D" and "Trap 0E"
handler along with a number of race conditions to eliminate
most of the crashes.
Unfortuanately, IBM also knew that their customers wanted Windows 95,
not because they WANTED Windows 95 but because they needed Windows 3.1
compatibility and the 3.1 crashes were killing their productivity.
Applications were programmed to "autosave" every 5-10 minutes to
prevent a disaster.
> > then you make everyone
> > sign an agreement
>
> At gunpoint?
At threat of bankruptcy. Which for the CEO is even worse than the
point of a gun. If you shoot a CEO, he dies, people remember him
as a great man who died too young, and he takes his place in history.
If you bankrupt the company of a CEO, he is forced to watch his
investors panic as the stock prices fall through the floor, he is
then forced to accept a hostile takeover bit at a fraction of the
company's original value, and he is finally put into some obscure
"President of Lunacy" position where he is kept quietly tucked away,
bound by a noncompete agreement, nondisclosure agreements, and
a golden parachute that's contingent on his sticking around for
many years without rocking the boat too much. He gets to go
from the "Alpha Male" to "Runt of the Litter", and stay there
for many many years. The most dramatic example was Steve Jobs,
who, after crossing Microsoft one too many times, was demoted
from CEO and given a position as "Chief Visionary" (it even sound
like Chief Lunatic). Investors wouldn't touch him with a 10 foot
pole (NeXT never got the funding it needed to be commercially viable).
It took him 10 years, including 3 years as "Interim CEO" to prove
that he really had the magic it took to make "Applesauce".
> No, you don't _make_ anyone do _anything_.
You give them a choice, one with minimal negative consequences and
marginal benifits, and the other with horrible consequences, huge
risk, and no possible benefits.
I have to respect Lou Gerstner and Michael Dell. Both of them
are taking a pretty big risk to make such a huge commitment to
Linux and UNIX. Carly Fiorina has been a bit more covert, as
has Michael Capellas. All have publicly supported Linux, which
is a direct challenge to Microsoft's draconian tactics.
> You offer them the rights to use a protocol,
While mentioning that since Microsoft is making this a keystone
feature of Windows 98 and Windows 2000, being EXCLUDED from that
protocol would mean that you would have to compete against a number
of Microsoft-backed competitors just to retain your market share.
And since Microsoft is putting so much of it's manpower into USB,
they really won't have the time to support any changes to your
current drivers. The old drivers **might** work O.K., but there
have been some changes to the kernel and we really haven't checked
that out (essentially guaranteeing that if you don't sign the
contract, your driver WON'T work). Legally, it's only extortion
if the perpetrator spells out the threat or the victim spells out
the threat and the perpetrator confirms with either a verbal
confirmation or a nondenial (changing the subject, trying to be
friendly, trying to make the victim feel guilty...) When you're
in prison, and your cell-mate has a knife to your throat and tells
you he just wants to be your "friend", you let him be your "friend".
> as long as they're willing not to
> disclose its operations, perhaps, but that's another matter.
The nondisclosure also spells out that if you DON'T sign the
contract/NDA, you will be excluded from the market. This isn't
protection of the various vendors, it's a veiled threat weilded
by Microsoft against all of the IHVs and OEMs.
Essentially, Microsoft is saying "I'm going to let you create a
protocol standard which Microsoft will control, and if you want
to use this standard, you must promise not to tell anyone about
this standard. Furthermore, if anyone does disclose anything
in this standard, they should be excluded from all future
discussions and upgrades to this standard. Bottom line "Spill
the beans, you're outta business".
> If the
> vendors really thought Linux and BeOS and suchlike were worth the
> effort, they'd reject proprietary protocols and adopt or establish
> open ones.
They tried. They published FireWire, with IEEE public specifications,
SCSI command set, 400 mb/s serial bus, and support for Linux, Windows
95, Windows NT, Windows 98, and Windows 2000, and because Microsoft
is leveraging billions to hype USB, FireWire can't reach the economies
of scale to provide low-cost storage devices to compete with USB.
The fact is that USBs vendors are struggling to get enough market
share to make USB worthwhile, but USB device makers know that
if they give too much support to FireWire, that Microsoft can
punish them in some nasty ways.
> If you have a case to make here, it's either against
> Linux for it's (real or perceived) lack of popularity, or at the
> hardware vendors for adopting closed protocols. Take it up with them.
You originally made the case that because Linux didn't support
Windows-only data formats (Office 2000) and Windows-only devices
(USB with complex command sets), that Linux couldn't survive.
I submit that Linux promotes a set of standards that is supportable
by Windows as well (nonexclusive), supports both environments, and
Linux supports a suite of devices which is also supported by most
versions of Windows.
> > > To be honest, I'm not sure they care about Linux...
> > > at least not in
> > > terms of where Linux is today.
> > Bill Gates and Steve Ballmer are acutely aware of Linux. It scares
> > the daylights out of them.
>
> Really? You've asked them, personally? I trust you have recordings.
Actually, you have the transcripts from the DOJ antitrust case. Are
you telling me they LIED on the Witness Stand?
> > Windows 95 was delayed nearly a year
>
> ...for various reasons, one of which might
> well include the one you claim.
Actually, I'd seen beta versions of Chicago (my NDA on this has
expired) at about the same time I saw Yddragasil.
Microsoft was also trying to capture some of the best features from
some of the X11 window managers (TWM, KDE, CDE, and OLVWM among them).
These two initiatives delayed Windows 95 by nearly a year (from stable
version of Chicago to stable version of Win95).
> > (which increased the cost of peripherals by about $20 each) using
>
> Yet I can get a 10Mbit PnP network adapter for $17, locally.
> According to you, without PnP, the vendor would actually have to pay
> me $3 for every one I took. Sorry, don't buy it.
Linux can use PnP LAN cards (though it does get confused by new
card signatures). You do pay extra for USB, WinModems, and, back
when PnP PCI first came out, you DID pay extra for PnP devices.
Today, PnP is pretty much cracked. The OEMs are being pretty generous
about disclosing compatibilites (our ethernet card is ne2k
compatible...), many are even providing drivers and driver settings.
> > a protocol which was contracturally protected from disclosure to
> > the Linux community.
>
> As well as anyone else, one would suspect.
Actually, they were specifically concerned about OS/2 and Linux.
OS/2 was still uncertain, and Microsoft wasn't telling IBM anything
until they agreed not to sell any more OS/2.
> > > Where Linux will be in 2-3 years
> > > depends on a lot of things,
> >
> > Keep in mind that we're really talking Linux/UNIX.
>
> Somehow I suspect that the commercial Unix variants regard Linux as
> _at best_ a competitor.
Sun considers Linux a Leverage tool. Linux has opened up a huge
number of markets that might have gone to NT, along with a number
of take-aways from NT. Sun often replaces existing Linux systems
when the web sites, mail servers, and groupware servers powered
by Linux create the justification for a UNIX server. Since Sun
has made it very easy to support porting of "Linux Applications"
(Actually Open Source applications which are compiled using an
open source compiler which runs on Solaris, AIX, and HP_UX), and
they understood the importance of Open Source to their own success,
(regularly interacting with customers directly via usenet newsgroups),
they were able to capitalize on some terrific opportunities created
by Linux. This is why Sun has about 1/3 of the sites on the Internet,
and over 1/2 the back-end servers used to support the web servers.
Sun also found that when they began providing direct support to
Blackstone (Java for Linux) to keep Linux up to date, that customers
were more willing to upgrade to newer versions of Java.
For developers who want to develop for a Solaris or AIX site,
Linux on the laptop or desktop is a very simple way to be able
to write code while propped up in your hotel or easy chair at home
instead of sitting on the office all day.
> > The combination of Linux and
> > UNIX as a solution that allows low cost
> > development, prototyping, and even mid-tier systems under Linux
> > that can easily be converted to BSD, Solaris, AIX, HP_UX, and
> > others.
>
> Yeah, but hiring a netadmin with a clue would also save you big
> bucks... at least, the way you've describe purchasing and maintenance
> is an absolute nightmare... and one totally unnecessary in the Windows
> world. Maybe Linux forces you into that sort of nonsense, but I'd be
> very surprised if it did. Step 1: shoot your existing net admin.
> Step 2: Hire one with a clue. Step 3: Get back to us in a year or
> two with the results.
Remember, you normally don't NEED to switch from Linux until you
get up into the OC5 feeds (155 megabits saturated). Many people
switch to commercial UNIX because they have some piece of commercial
software (MQSI, WebLogic,...) that is only available on commercial
UNIX, at which point, they just start rolling more and more of the
services running on PC-sized Linux boxes onto the bigger Sun boxes.
Eventually, you see from the PC usage meter (not from down-time or
response time) that the CPU is at near 80% (I've seen UNIX systems
run quite successfully with 400%), and they get another machine to
spread the load.
Some folks stay with Linux and go into beowulf clusters. Even trivial
clusters can produce a lot of horsepower for very little cost.
> > We're also talking Open Source vs
>
> Whoopee.
>
> If I want a programmer to write an application, I want him to write
> the _application_, not fart around custom tailoring OS code.
Precisely. If your application programmer has access to libraries
and programs that have been field tested under extreme loads, and
has the ability to step into the subroutine to see where he went
wrong (usually a misunderstood pointer to pointer usage such as
some of the termcap() functions, they can sort out the problem
in a few hours instead of waiting from some Microsoft guru to take
a look at their code and tell them to redesign that whole section
of the project.
> If I hire a network admin, it's to administer
> the network, not redo OS code.
Yes, but if you have a server that's failing more than a few times
in a month, it's nice to know that you can get it fixed.
By the way, just because the administrator has the source code,
doesn't necessarily mean that he's the one who will fix it. When
you drive your care, you don't need the blueprints, but when your
mechanic is replacing a water pump, you sure are glad he has that
old Chilton's manual.
> If I hire documentation people, clerical staff, accountants,
> managerial staff, it's all to do jobs having exactly nothing to do
> with farting around with OS code.
Yup. And if something goes wrong with the accounting system,
you really want someone who can get in there and figure out what
when wrong, and make sure it never happens again.
> So what does having an open-source
> OS buy me? Sweet f a, unless I'm an OS geek. Fine, if I'm an OS
> geek, I'll go work for Microsoft, or Be, or RedHat.
Nope. It means that BSOD, that locked up server, or that slow
performance on a query, can be fixed so that it never fails again.
> > Linux has already done a remarkably good job of
> > creating a "user desktop" appliance in remarkably
> > short time.
>
> I've used it; it's a joke. Oh, it's a *hell* of a lot better than it
> was a couple years back... but it's got a hell of a long way to go,
> especially considering the alternatives.
That's especially funny. With Windows 9x or NT, there are NO
alternatives. Even Active Desktop is an add-on, bundled with features
that make most security officers, network administrators, and
project managers cringe (just what we needed, streaming data to
destract users all during the work day, load up the network, and
force-feed viruses without even asking for permission to load).
> > by many to be a viable alternative to Windows. Many
> > people actually prefer Linux because it is so much more
> > flexible, reliable, secure, and stable.
>
> Yes, many do... but look at who they are.
> Did your folks buy a computer?
Yup - My dad's first was a commodore 64.
> Was it a Linux box?
He runs Linux sometimes. My little brother (a mechanic)
loves Linux. He uses it to play games and tinker with his
web site.
My son likes Linux, he's been creating a web site too.
My daughter likes Windows, but only because her brother keeps
hogging the Linux box.
My girl-friend can't tell the difference between Linux and Windows,
she just points the browser and clicks. She uses netscape
communicator too.
My ex-wife hates all computers, especially Windows.
> Probably not. Check most of the
> offices around you; do they buy machines? Are they usually Linux
> machines? Probably not.
But when I mention Linux, 1/4 have it at home, 1/5 have it on their
workstation, and 1/2 want to know more about it. Granted, this is
the techie area (admins, programmers, and all-round geeks).
On the train, I get stock brokers running it on their laptop,
and accountants running it at home (home business).
> What you find is lots of so-called "power users" going for Linux, for
> those reasons. Sure, no argument, it *does* have advantages.
Yes, reliability, security, stability, functionality, and flexibility.
For the price of Windows 95, you get functionality of BOTH NT and UNI
> However, until it's as easy for everyone in the office to use as
> Windows, and that includes interoperability with the other offices in
> the organization, which are probably already Windows boxes, well,
> sorry.
But you loose:
flexibility - ability to configure what's needed for the job.
security - ability to know the nature and content of your information.
> > Linux provides much better compliance with existing IETF standards.
IETF is Internet Engineering Task Force (Internet RFCs).
> > Linux supports a number of formats and standards that can be easily
> > implemented on Micrsoft Windows (95,98,NT, & 2000).
>
> Sure - if anyone in the Windows world cares. Do they?
We're talking about RTF, HTML, GIF, JPEG, CGM, and EPS.
No big huge software replacement, just choose the "save as"
option when you first start editing the document. Now you have
pretty documents that everybody can read.
> > Actually, KDE combined with GNOME and Java have created a user
> > oriented environment that offers features Windows can't even
> approach.
>
> KDE _looks_ nice enough, it looks like it solves a lot of the
> usability issues... but it has _got_ to do something about the speed.
When choosing hardware for Linux, chips that support accellerators
perform much better than those that use the SVGA drivers.
Ever run Windows NT on a generic VGA card? It's pretty slow too.
> > The ability to access remote workstations and servers using GUI
> > interfaces (application servers that offer XClient software).
>
> Ooh. Can't get that out of the box with Windows, but there are
> certainly X Servers available. Check Attachmate's KEA! X product, for
> example.
Problem there is you don't have X clients. I can use exceed,
hummingbird, or a number of X11 servers to get to UNIX and Linux
systems, but I'm still stuck in windows. I can't get to Windows
boxes from either Windows or X11 boxes.
> Also look into TSE... funny, I can log into a Terminal Server with a
> GUI client, run my programs, do my work. No, it's not X... but so
> what? It's here, it works.
Most of these give 1 user access to each application. If two people
want to run Word, too bad.
> > They will rely on NDAs and non-compete
> > contracts to prevent the spread
> > of Linux as much as possible,
>
> Hmm... I sign an NDA with MS, and this somehow prevents Linux
> spreading? What, the magic OS pixies start randomly deleting Linux
> kernel code?
Nope. OEMS want to put Linux and Windows on the same machine.
Ironically, Microsoft WILL let a "Linux" company like VA Linux do that,
but not a "PC" company.
--
Rex Ballard - Open Source Advocate, Internet
I/T Architect, MIS Director
http://www.open4success.com
Linux - 40 million satisfied users worldwide
and growing at over 5%/month! (recalibrated 7/2/00)
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: "Stuart Fox" <[EMAIL PROTECTED]>
Subject: Re: Samba vs NT, which gives best PCs / Server Performance?
Date: Tue, 25 Jul 2000 08:10:28 +0100
"Woofbert" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> In article <[EMAIL PROTECTED]>, Craig Kelley
> <[EMAIL PROTECTED]> wrote:
>
> As I've been saying all along, a competent network administrator can
> support a heterogeneous population of operating systems ... while NT
> administrators whine and moan about having to support Macintosh.
>
> --
Mac support isn't hard to implement under NT.
------------------------------
** 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
******************************
