Vista Will Exterminate Desktop Linux Once And For All.

[arachnid]

<snip>

I have to go with Kerry Brown. He is showing some business sense.

It's cost that counts. Two OSs cost more to support than one.

Why should it cost any more to support, say, 40,000 Windows machines and
40,000 Linux machines, than it does to support 80,000 Windows machines?

What if the Linux machines require *less* tech support than Windows?

What if, as with many lower-end PC's, the manufacturer's "tech support" is
limited to reinstalling the OS from a rescue CD?

And that's hard for a manufacturer to swallow, unless there's a higher
markup to justify the second (Linux) OS.

OTOH the manufactuer has to swallow the cost of all those Windows
licenses, whereas they can install Linux for free.

In the mainstream PC market, low-end products dominate. Margins are tight,
and makers have to work hard to stay profitable. Supporting a second OS in
your low-margin products is hard.

Contrast this with the server market. Where margins are greater, Linux can
get in more easily.

Because it's free, it also gets in where margins are smaller. )

Further, Linux provides a natural migration path for Unix server shops. For
the customer, Unix (whether HP, Sun, IBM, or other) is MUCH more expensive
than Linux, both to own and to maintain. That's where Linux slides in
easily. (HP will soon drop HP-UX. IBM is already pushing Linux more than
AIX. I wonder when Sun will drop Solaris?) Lots of growth potential.

But, domestically, I see Linux growing only slowly on the desktop, due to a
high level of dependence on existing Windows software.

I won't deny that there are some Windows applications that Linux doesn't
have any satisfactory equivalent for just yet, but there are enough
mature Linux applications to satisfy the needs of -most- users. I could
sit your average home MS-Windows user down to a pre-installed Ubuntu Linux
system and have them using it comfortably in less than an hour. Any
problems they have after that aren't likely to be any more frequent or
severe than if they were using Windows.

 

[Kerry Brown]

arachnid wrote:

I won't deny that there are some Windows applications that Linux
doesn't have any satisfactory equivalent for just yet, but there are
enough mature Linux applications to satisfy the needs of -most-
users. I could sit your average home MS-Windows user down to a
pre-installed Ubuntu Linux system and have them using it comfortably
in less than an hour. Any problems they have after that aren't likely
to be any more frequent or severe than if they were using Windows.

Here's a scenario for you.

The average user buys a pc with an alternative OS on it. He happily connects
to the Internet, surfs, uses an instant messaging client to talk to his
buddies, burns some music to a CD, and is generally pretty happy with his
purchase. After a few days he gets an email from a friend with some great
pictures he'd like to print out. He wanders down to his local Wal-Mart and
buys a $79.99 printer. When he gets home he can't figure out why the printer
isn't working. He calls the printer manufacturer's support. They tell him
they don't support his alternative OS and he should contact the manufacturer
of his pc. He phones them and waits for 1/2 an hour while he is being
transferred to the alternative OS support group. They walk him through some
troubleshooting steps and finally figure out he didn't purchase a USB cable
with his printer. He's more than a little upset so he decides not to go back
to Wal-Mart as they could have told him he needed a cable. He drives a
little further this time to a major name electronics store as they must be
computer experts. He explains his problems, the salesman is very sympathetic
and sells him a super duper gold plated USB cable that has purple see
through insulation for only $49.99 giving him a deal ($10.00 off) because he
signed up for their company email newsletter. He gets home, hooks up the USB
cable, fires up Open Office and trys to print but gets nothing. He phones
the printer manufacturer only to be told once again they don't support his
alternative OS. He phones the computer manufacturer this time after waiting
1/2 an hour he gets a different rep who informs him that his XYZ printer
isn't compatible with his alternative OS but he can sell him an ABC printer
which will work just fine. How satisfied is this customer? Is he likely to
buy a computer with an alternative OS next time? Is he likely to buy the
same brand of computer next time? How much extra time did the OEM have to
spend with him (hint: time is money). How much did it cost the OEM to set
up, hire, and pay the extra support staff for the alternate OS?

"Any problems they have after that aren't likely to be any more frequent or
severe than if they were using Windows." Indeed.

We haven't even started to discuss other things that might crop up. What if
the purchase is a laptop but is very low end, doesn't have wireless, and the
customer buys a USB wireless dongle? What if the user needs to use a
specific LOB application for his job? Or even to setup a VPN to his
corporate network. Many of these scenarios will involve support calls to the
OEM manufacturer. Even if the call lasts five minutes and consists of "we
don't support that" it costs the OEM money to have the staff available to
say that. Even if they don't get one extra call over supporting one OS and
don't hire any extra staff they still have to train the existing staff on
the new OS. Everything has a cost. Unless there is a payoff OEMs have no
interest in spending more money or changing the status quo. It may not be
the way we wish things to be but it is the reality.

 

[cc]

When Vista hits the streets Linux on the desktop will become a
has-been. The truth is that Linux has had 15 years of Microsoft's
blunders to gain the confidence of enough destkop users to make some
kind of impact on the desktop and it has failed to do so.

Microsoft Vista is generating excitement amongst the Windows users,
which account for approximately 95 percent of the desktop users.
PC Magazine, PCWorld and all the other main stream magazines devoted to
PC computing are filled with articles concerning Vista.

http://www.pcworld.com/
http://www.pcmag.com/

Jesus Christ who reads those dork mags. Is Vista really generating
excitement? If so we have a problem of unparalleled dorkitude on our
hands. Which losers are going to run out at 12AM to be the first to get
their copy of Windows? Don't laugh linux users, you people are just as
bad, but at least you stay in your homes.

 

[arachnid]

arachnid wrote:



Here's a scenario for you.

You've only shown that if one tries, it's possible to concoct convenient
scenarios out of thin air to prove your point. I could create one about a
user that buys a Windows machine and keeps getting infected by spyware. Or
how about the user who moves his music to the new machine and it won't
work because he needs the DRM encryption key that's on the old machine he
just wiped and donated to charity, and now he's tying up tech support
demanding the OEM make his $500 worth of music work? How about the user
who buys a computer that doesn't come with Windows and expects to be able
to install OEM Windows from his previous computer? How about a program
installation munging the Registry or even just changing file associations?

I can cook up some more if you want. )

"Any problems they have after that aren't likely to be any more frequent
or severe than if they were using Windows." Indeed.

We haven't even started to discuss other things that might crop up.

What if <snip>

What if someone buys a printer that only comes with Vista drivers and
wants to use it on their Windows XP machine? What if they buy a new Vista
machine, don't like Vista, and want to install Windows XP - only XP
doesn't support some of the newer Vista-compatible hardware?

(That last is taken from a similar personal experience with Win98SE and
XP. I tried to install Windows 98SE on this 64-bit notebook and the
installer crashed. So I installed 98SE onto the hard drive using another
system, prepped it for a system move - and it still crashed)

Unless there is a payoff OEMs have no interest in spending more money or
changing the status quo. It may not be the way we wish things to be but
it is the reality.

The reality is that manufacturers have to pay a licensing fee for Windows
that, for a low-end laptop, can amount to as much as 25% of the *retail*
cost of the laptop, whereas they pay nothing for Linux. Think that might
offset a few extra tech support calls?

In my experience fixing friends' systems, the majority of Windows problems
are caused by spyware and adware that probably got in via IE or Outlook
Express. We don't have that problem in Linux. So who is to say that all
the problems you concocted, even if they really occurred, wouldn't be
offset by the savings on licenses plus an end to spyware/adware-related
problems?

 

[arachnid]

Jesus Christ who reads those dork mags. Is Vista really generating
excitement? If so we have a problem of unparalleled dorkitude on our
hands. Which losers are going to run out at 12AM to be the first to get
their copy of Windows? Don't laugh linux users, you people are just as
bad, but at least you stay in your homes.

Yeah, I'm really eager to see the new version of Ubuntu due out this
month. But instead of trying to sleep on the sidewalk in front of CompUSA,
I think I'll just drop a DVD blank in the burner and run a little script
to download and burn the ISO while I sleep.

 

[cc]

That's probably the biggest weakness of the Linux business model. What OEM
wants to spend 4-10x the effort to make sure dozens of distros work with
their hardware when they can do it once, with one OS?

Remember, we're talking about companies that have to support their end
users. They have 1-3 year on-site waranties that cover both the OS and
Hardware. Dell, for example, sells a lot of 3 year support contracts.

prove it. what's a lot?

So, you might think, pick one distro and be done with it? But which one?
Linux users are fickle, and distro loyalties seem to change overnight. One
month it's Red Hat, then it's SUSE, then it's Debian, than it's Ubuntu. It
all seems to depend on who leapfrogs whom, until the next one does the
same.

On the server, it's a little easier. There are only a handful of
"enterprise" server distro's, and they seem pretty stable. On the desktop
it's a totally different story. With Linux in such constant state of flux,
it's impossible for any vendor to choose just one and stick with it, at
last not if they don't want to lose most of their customers in 6 months.

prove it.

And even if, by some miracle, the distro they choose is stable, and stays
somewhat popular, it's still only going to be somewhere between 15 and 30%
of the customers will want that distro, so you're stuck supporting multiple
ones.

That means training your support staff to handle multiple distro's.
Especially the front line grunts who aren't computer experts. They have to
know the difference between apt, and emerge, and rpm. They have to know
how configuration files are stored in different places on different
distro's.

The fact of the matter is, Linux is too volatile for standard OEM support,

prove it.

 

[cc]

Yeah, I'm really eager to see the new version of Ubuntu due out this
month. But instead of trying to sleep on the sidewalk in front of CompUSA,
I think I'll just drop a DVD blank in the burner and run a little script
to download and burn the ISO while I sleep.

Yeah, I care.

 

[Erik Funkenbusch]

More Erik Fudmeister lies. 4-10x the effort indeed.

There's only a few major ones, Erik. If they work, then most all will
work.

Not true at all.

The single biggest example would be when the hardware requires a device
driver that isn't provided by the stock kernel. The vendor then has to
create packages for all the possible versions of Linux they want to
support. Yes, there will be some cases where a package will work with
multiple distro's, they would have to verify that on each distro to support
it. You can't just assume that any RPM package will work correctly on any
system that uses RPM, because files are stored in different locations in
different distro's and different distro's kernels are compiled with
different options.

 

[cc]

Not true at all.

The single biggest example would be when the hardware requires a device
driver that isn't provided by the stock kernel. The vendor then has to
create packages for all the possible versions of Linux they want to
support. Yes, there will be some cases where a package will work with
multiple distro's, they would have to verify that on each distro to support
it. You can't just assume that any RPM package will work correctly on any
system that uses RPM, because files are stored in different locations in
different distro's and different distro's kernels are compiled with
different options.

This is totally and completely stupid.

 

[Kerry Brown]

You've only shown that if one tries, it's possible to concoct
convenient scenarios out of thin air to prove your point. I could
create one about a user that buys a Windows machine and keeps getting
infected by spyware. Or how about the user who moves his music to the
new machine and it won't work because he needs the DRM encryption key
that's on the old machine he just wiped and donated to charity, and
now he's tying up tech support demanding the OEM make his $500 worth
of music work? How about the user who buys a computer that doesn't
come with Windows and expects to be able to install OEM Windows from
his previous computer? How about a program installation munging the
Registry or even just changing file associations?

I can cook up some more if you want. )


What if someone buys a printer that only comes with Vista drivers and
wants to use it on their Windows XP machine? What if they buy a new
Vista machine, don't like Vista, and want to install Windows XP -
only XP doesn't support some of the newer Vista-compatible hardware?

(That last is taken from a similar personal experience with Win98SE
and XP. I tried to install Windows 98SE on this 64-bit notebook and
the installer crashed. So I installed 98SE onto the hard drive using
another system, prepped it for a system move - and it still crashed)


The reality is that manufacturers have to pay a licensing fee for
Windows that, for a low-end laptop, can amount to as much as 25% of
the *retail* cost of the laptop, whereas they pay nothing for Linux.
Think that might offset a few extra tech support calls?

In my experience fixing friends' systems, the majority of Windows
problems are caused by spyware and adware that probably got in via IE
or Outlook Express. We don't have that problem in Linux. So who is to
say that all the problems you concocted, even if they really
occurred, wouldn't be offset by the savings on licenses plus an end
to spyware/adware-related problems?

The license is a small cost overall in the manufacturing process. For a
small OEM like myself it is a factor. For someone making 10,000 computers a
day in factories worth hundreds of millions of dollars it is not that
significant. Support is a more significant factor for them. Both from the OS
developer and that they supply to their customers. The spyware/adware
related support calls is a problem I hadn't thought of. You're right
reducing those calls would be a cost savings. I knew that my mythical
scenario could be easily argued against. It was meant to be a bit of humour
if nothing else. I would actually like to see someone come out with an OEM
computer with an option for an alternative OS to Windows. I have been trying
to sell Linux based computers for years. It is a very hard sell. If someone
started a marketing campaign I am sure I could cash in on it. As I said in
another post Apple is gaining some market share. Their Mac vs PC commercials
have been very effective. That's what it will take, good marketing. I can't
see any of the current big players doing it. They would rather have a closed
market with the status quo. It would take a smaller player with deep pockets
and a commitment to make it work.

 

[arachnid]

Not true at all.

The single biggest example would be when the hardware requires a device
driver that isn't provided by the stock kernel. The vendor then has to
create packages for all the possible versions of Linux they want to
support.

OEM's don't choose the hardware first and then write drivers to make
the OS fit the hardware. They buy hardware that works with the OS they
intend to put on the system. If they want the hardware really bad and it
doesn't work with the OS, it's usually the seller - the hardware
manufacturer - that has to develop the drivers. No drivers, no sale.

Yes, there will be some cases where a package will work with
multiple distro's, they would have to verify that on each distro to support
it. You can't just assume that any RPM package will work correctly on any
system that uses RPM, because files are stored in different locations in
different distro's and different distro's kernels are compiled with
different options.

Just give us the specs. OSS developers will handle the rest. We actually
prefer it this way because OSS developers are more familiar with OSS
software standards and better at writing Linux drivers, and having our own
source means we're no longer dependent on the hardware manufacturers. So,
for example, when we moved to 64 bits we were able to recompile our device
drivers whereas Windows users were stuck in 32-bit land because
the peripheral manufacturers didn't see any percentage in developing
64-bit drivers for hardware they didn't make anymore.

 

[arachnid]

he license is a small cost overall in the manufacturing process. For a
small OEM like myself it is a factor. For someone making 10,000
computers a day in factories worth hundreds of millions of dollars it is
not that significant.

Given the cutthroat margins in the computer industry, 10%-25% of the cost
of making the computer is not "insignificant" no matter how many millions
of dollars of business you do.

Support is a more significant factor for them. Both from the OS
developer and that they supply to their customers. The spyware/adware
related support calls is a problem I hadn't thought of. You're right
reducing those calls would be a cost savings. I knew that my mythical
scenario could be easily argued against. It was meant to be a bit of
humour if nothing else. I would actually like to see someone come out
with an OEM computer with an option for an alternative OS to Windows. I
have been trying to sell Linux based computers for years. It is a very
hard sell.

That I'll agree to. No matter how good Linux gets, it's going to take a
big manufacturer with some staying power to persuade people to give up the
devil they know.

If someone started a marketing campaign I am sure I could cash in on it.

Everyone's waiting around for someone else to be first.

As I said in another post Apple is gaining some market share. Their Mac
vs PC commercials have been very effective. That's what it will take,
good marketing. I can't see any of the current big players doing it.

That's very likely to be Mark Shuttleworth. He founded Thawte and sold it
for over $500 million so he has both the deep pockets and the business
skills. Judging by Ubuntu he also groks "user-friendly" better than
anyone else in the Linux world, which may be the most important thing of
all.

Whoever the OEM is, I'm certain the distro will be Ubuntu - and I think
Shuttleworth will do whatever advertising is necessary to get Ubuntu
accepted by the market. In fact it seems he's already doing some
marketing - he's mailing out free CD's to anyone who asks, and Ubuntu
billboards have been appearing in southern California.

They would rather have a closed market with the status quo. It would
take a smaller player with deep pockets and a commitment to make it
work.

The big players are being dragged in kicking and screaming because
governments and large businesses are beginning to adopt Linux. Losing a
bid for 5,000 machines because they're not linux-compatible tends to get
peoples attention, so most OEMs now have Linux compatible machines. They
don't advertise them as such because they're afraid Microsoft will punish
them with higher license fees, but they're available if a business
specifically asks for them.

 

[Paul Hovnanian P.E.]

When Vista hits the streets Linux on the desktop will become a
has-been. The truth is that Linux has had 15 years of Microsoft's
blunders to gain the confidence of enough destkop users to make some
kind of impact on the desktop and it has failed to do so.

There isn't that much difference between W2K/XP and Vista to make much
of a difference, market-wise. The event that affects Linux on the
desktop is each time Microsoft discontinues support for one of their
older OSs. Each time they do, the group of holdout users has to choose
between the pain of switching to Linux or the pain of switching to the
latest Windows.

The Window pain (sorry) is getting worse every time.

Microsoft Vista is generating excitement amongst the Windows users,
which account for approximately 95 percent of the desktop users.

Yes. Maybe THIS time, they'll get it right. Its like buying a Powerball
ticket.

PC Magazine, PCWorld and all the other main stream magazines devoted to
PC computing are filled with articles concerning Vista.

Filled with Microsoft ads too.

 

[Tim Smith]

support. Yes, there will be some cases where a package will work with
multiple distro's, they would have to verify that on each distro to support
it. You can't just assume that any RPM package will work correctly on any
system that uses RPM, because files are stored in different locations in
different distro's and different distro's kernels are compiled with
different options.

I wonder if this is mostly an RPM problem? In the RPM world, you've got
Redhat and SuSe, who have gone in different directions. For example, I
was trying to update some software on a SuSE box, and the RPMs I found
on the net apparently were for distributions that followed the Redhat
way. They depended on packages that don't exist in SuSE.

All of the deb distributions I can think of offhand are built on Debian
itself, either directly (Ubuntu) or indirectly (e.g., ones that are
built on Ubunutu). That should greatly lessen this package problem.

 

[miss-information]

...I would actually like to see someone come out with an OEM computer
with an option for an alternative OS to Windows.

FreeDOS?
http://www.dell.com/content/topics/segtopic.aspx/e510_nseries?c=us&cs=19&l=en&s=dhs&~ck=mn

mi

 

[altheim]

arachnid wrote:



Here's a scenario for you.

The average user buys a pc with an alternative OS on it. He happily
connects to the Internet, surfs, uses an instant messaging client to talk
to his buddies, burns some music to a CD, and is generally pretty happy
with his purchase. After a few days he gets an email from a friend with
some great pictures he'd like to print out. He wanders down to his local
Wal-Mart and buys a $79.99 printer. When he gets home he can't figure out
why the printer isn't working. He calls the printer manufacturer's
support. They tell him they don't support his alternative OS and he should
contact the manufacturer of his pc. He phones them and waits for 1/2 an
hour while he is being transferred to the alternative OS support group.
They walk him through some troubleshooting steps and finally figure out he
didn't purchase a USB cable with his printer. He's more than a little
upset so he decides not to go back to Wal-Mart as they could have told him
he needed a cable. He drives a little further this time to a major name
electronics store as they must be computer experts. He explains his
problems, the salesman is very sympathetic and sells him a super duper
gold plated USB cable that has purple see through insulation for only
$49.99 giving him a deal ($10.00 off) because he signed up for their
company email newsletter. He gets home, hooks up the USB cable, fires up
Open Office and trys to print but gets nothing. He phones the printer
manufacturer only to be told once again they don't support his alternative
OS. He phones the computer manufacturer this time after waiting 1/2 an
hour he gets a different rep who informs him that his XYZ printer isn't
compatible with his alternative OS but he can sell him an ABC printer
which will work just fine. How satisfied is this customer? Is he likely to
buy a computer with an alternative OS next time? Is he likely to buy the
same brand of computer next time? How much extra time did the OEM have to
spend with him (hint: time is money). How much did it cost the OEM to set
up, hire, and pay the extra support staff for the alternate OS?

"Any problems they have after that aren't likely to be any more frequent
or severe than if they were using Windows." Indeed.

We haven't even started to discuss other things that might crop up. What
if the purchase is a laptop but is very low end, doesn't have wireless,
and the customer buys a USB wireless dongle? What if the user needs to use
a specific LOB application for his job? Or even to setup a VPN to his
corporate network. Many of these scenarios will involve support calls to
the OEM manufacturer. Even if the call lasts five minutes and consists of
"we don't support that" it costs the OEM money to have the staff available
to say that. Even if they don't get one extra call over supporting one OS
and don't hire any extra staff they still have to train the existing staff
on the new OS. Everything has a cost. Unless there is a payoff OEMs have
no interest in spending more money or changing the status quo. It may not
be the way we wish things to be but it is the reality.

I hate to do this to you but here's a quote from one of your earlier
posts (yesterday 9/10 in fact): " I personally would like to see more
diversity in the OS marketplace. Everyone would benefit in the long
run. Competition improves a product."

Your scenario, which I agree with BTW, demonstrates entirely the
opposite. We were not discussing lawn mowers at the time you made
the above observation, or furniture or TV sets or other standalone
hardware where a choice of product is desirable, but about operating
systems where users, and the various bits of connected hardware they
use, are dependant upon it to provide a standard means of
communicating and interacting with each other.

Can you explain? Have you had a change of heart?

 

[Erik Funkenbusch]

OEM's don't choose the hardware first and then write drivers to make
the OS fit the hardware. They buy hardware that works with the OS they
intend to put on the system. If they want the hardware really bad and it
doesn't work with the OS, it's usually the seller - the hardware
manufacturer - that has to develop the drivers. No drivers, no sale.

That means they may well have seperate hardware for Linux and Windows.
More cost. But I was thinking more along the lines of companies like Dell
that make their own hardware (or license it from other comapnies) such as
their PERC RAID controllers. Other vendors have similar custom hardware.

Just give us the specs. OSS developers will handle the rest.

Wrong answer. We're talking about OEM's, who have support contracts with
their custoemrs. They're not going to rely on someone they don't also have
contracts with to supply code in case of new hardware and/or fixes.
Further, even if they did just supply the specs, there is no guarantee
anyone will write the software they need in any reasonable time frame.
Yes, they can hire their own programmers to do it, but that's exactly the
point. More cost.

We actually
prefer it this way because OSS developers are more familiar with OSS
software standards and better at writing Linux drivers, and having our own
source means we're no longer dependent on the hardware manufacturers.

That's just it, the hardware manufacturers need some dependance, because
they need to know when software will be ready, and guarantee it follows
THEIR standards.

This comment just proves how little you understand about the OEM market.

 

[Hadron Quark]

I hate to do this to you but here's a quote from one of your earlier
posts (yesterday 9/10 in fact): " I personally would like to see more
diversity in the OS marketplace. Everyone would benefit in the long
run. Competition improves a product."

There is a big difference from wanting competition, to having half assed competition.

His scenario is quite common : it happened to me with my printer. I'm
more careful now. Unfortunately I believed the COLA gang when they said
"it all just works". Now the mantra is "check before you buy".

Your scenario, which I agree with BTW, demonstrates entirely the
opposite. We were not discussing lawn mowers at the time you made
the above observation, or furniture or TV sets or other standalone
hardware where a choice of product is desirable, but about operating
systems where users, and the various bits of connected hardware they
use, are dependant upon it to provide a standard means of
communicating and interacting with each other.

Can you explain? Have you had a change of heart?

There is no change of heart : he wants competition.

 

[Erik Funkenbusch]

I wonder if this is mostly an RPM problem? In the RPM world, you've got
Redhat and SuSe, who have gone in different directions. For example, I
was trying to update some software on a SuSE box, and the RPMs I found
on the net apparently were for distributions that followed the Redhat
way. They depended on packages that don't exist in SuSE.

All of the deb distributions I can think of offhand are built on Debian
itself, either directly (Ubuntu) or indirectly (e.g., ones that are
built on Ubunutu). That should greatly lessen this package problem.

While that's true of Ubuntu, that's because Debian has a huge
infrastructure in place. There are a lot of RPM based distros, though.
SUSE, Red Hat, Mandrake, etc.. not to mention the one-offs, like Gentoo,
Gobo, Linspire, Xandros, etc...

 

[Erik Funkenbusch]

Given the cutthroat margins in the computer industry, 10%-25% of the cost
of making the computer is not "insignificant" no matter how many millions
of dollars of business you do.

You're making the classic mistake. Suppose Vendor A offers a model without
an OS. By doing this, they can cut their price and undercut their
competition. However the competition will react and it won't take long to
be back at that 6% margin again because all yoru competitors have undercut
you to the bone.

Regardless of what products you sell, there will always be a very thin
margin, so dropping Windows (or any other OS for that matter) won't give
them bigger margins.