casioculture@xxxxxxxxx wrote:
> Oh man. It'll take forever.
>
> "For the desktop, market share in mature markets will be single digits
> in five years; in emerging markets it will be in the 15 percent to 20
> percent range."
>
> http://www.internetnews.com/dev-news/article.php/3627061
<quote>
Hohndel predicted that the embedded market will be 80 percent Linux in
the next five years.
For the desktop, market share in mature markets will be single digits
in five years; in emerging markets it will be in the 15 percent to 20
percent range.
"Adoption of the Linux desktop is more likely in emerging markets where
there is no legacy," Hohndel said.
Raymond got riled up as he proclaimed what he thought was necessary to
be done for desktop Linux to be successful.
"We need to do whatever compromise is necessary to get full multimedia
capability on Linux so non-technical users don't dismiss us out of
hand," Raymond shouted.
A somewhat more relaxed DiBona advised the audience to tell people to
use Mozilla Firefox on their desktops.
"Develop for the Web," DiBona said. "People can switch to Web
applications from their desktop more easily."
</quote>
This is a pretty good picture.
Emebedded Linux is all over the place. If you lump embedded linux with
the rest of the market, there are actually more Linux devices than
Windows devices being sold every year. Within a year or two there may
be more *nix devices deployed than Windows devices, and most people
won't even know they are using Linux. Perhaps if Linus insisted that
Linux trademark and logo be used on all devices that use Linux, that
would change.
Most people spend as much as 80% of their time interacting with web
applications hosted by or connected to Linux/Unix systems.
And yet, the Windows "Web" APIs limit the user to less than 5% of the
power available through Unix systems.
AJAX and OSS APIs are making it much easier to get the full
capabilities of Linux regardless of the desktop. In many ways, OSS is
invading the desktop. Firefox, OpenOffice, and so on.
I was surprised at the "single digit" rate. I'm also curious as to
which metric he was referring. Perhaps he was talking about the number
of machines sold with Linux (only) preinstalled.
I also agree with Eric Raymond, the "open-source only" mentality has
given way to a more pragmatic approach to coexistence with commercial
software.
The one thing that always comes to mind, is the experience of being
actively involved (and even ridiculed for) in the commercialization of
the Internet. The "smart money" was backing Netware IPX/SPX, and
Microsoft NetBIOS. The TCP/IP protocol was rarely touted by vendors
and usually reluctantly shown only when a customer insisted.
But then one day, Mosaic 2.0 came out and over 200,000 copies were
downloaded in less than a week. Counts using cookies (just added to
Mosaic 2.0) showed that around 2 million people were using the Internet
for web access. That was more than AOL, Prodigy, and Compuserve at the
time. Even more interesting, market metrics showed that the Internet
user base was doubling in size every 4 months. Within 16 months, that
2 million users had become 32 million users. That was more than all
other services COMBINED.
If you weren't actively involved, you would never have predicted this
kind of phenominal growth. You would never have predicted the impact
that the "Internet" would have on our culture.
But if you knew the problems of getting connected in 1991, and you
understood the network of BBS operators who had switched to TCP/IP in
1992 and 1993, and you understood the ease of connecting made possible
by trumpet winsock and Mosaic at the end of 1993, you knew that by
1994, the numbers would be crazy. And if you knew that MCI had agreed
to provide a commercial TCP/IP backbone, and connect that to corporate
users, you knew that the corporate market would explode.
So here we are, 12 years later, and there are some fundamental triggers
that appear to be about to trigger a similar growth in Linux user
base..
Knoppix - no need to do a full install or repartition the hard drive,
just plug in the CD and a USB drive, and you have Linux.
VMWare Player - Suddenly, any windows user can install a fully
functional, fully configured Linux system on their desktop, without
losing Windows XP functionality.
AMD-64 - Microsoft delayed the deployment of 32 bit PCs for almost 10
years, and by the time NT was finally stable (NT 4.0 with SP3), Linux
already had support for 64 bit Alpha, MIPS, Ultrasparc, and soon after,
PowerPC. When AMD came out with a 32/64 bit chip, HP decided not to
wait for Microsoft. To establish a market, they sold the machine with
Windows, but tested and supported it with Linux, by choosing Linux
oriented hardware components and providing drivers for new hardware.
Broadband Internet - Instead of going to retail stores, which
irregularly stocked Linux inventory and frequently had out of date
version, Linux could be downloaded from distributors over cable-modems
and DSL lines at rates up to 1 megabyte per second. A CD sized file
could be downloaded in an hour, a DVD sized file could be downloaded in
about 4 hours.
Linux appliances - in the datacenter and in the home, Linux was
establishing itself as a very strong player. It was reliable, secure,
and barely needed attention. The desktop distributions were loaded to
the gills with OSS applications.
The final driver may end up being Micrososft itself:
Windows XP -force feed: When XP was released, Micrososft came to it's
corporate customers and demand that they immediately sign a contract
committing them to triple the monthly payments. If they refused, or
even delayed, they were told that they would be "on their own", getting
no support of any kind. Microsoft didn't tell them that everyone would
be getting update services. As a result of this tactic, nearly all
major Microsoft customers have made plans to migrate to Linux and have
implemented steps toward that transition, without alerting Microsoft.
The OEMs have notices this trend, and have converted nearly all of
their lines to "Sold with Windows, Ready for Linux" (SWW-RFL). The key
feature, one which has lead to increased demand, being RFL.
Longhorn/Vista: Microsoft almost lost control of the market when they
released NT. It needed expensive hardware upgrades including increased
memory and hard drive space. Even though it was announced an 1992 and
was to be delivered in 1993, it wasn't even stable enough for general
use until late 1997, and didn't meet it's original promises until 2003
(windows XP). If a contractor promised to build you a house in a year,
and didn't deliver a livable functional house until 10 years later,
would you hire him again?
In 1993, Linux was stable, secure, and functional, but it as hard to
install. You had to purchase equipment from a "compatibility list"
that could be printed on the back of a VCR sized box. You had to
configure it yourself, manually. If you tried to push the performance
of the hardware too far, like pushing the resolution of the monitor, it
could catch fire. It had most of the features of a Sun SparcStation,
and even looked like one from a distance.
In 2006, Linux is stable, secure, functional, and can be installed on
about 80% of the computers sold in the last 2 years. A broad spectrum
of applications is now available, both in Open Source and in Commercial
forms. Most new applicatiotns are now offered in Windows and Linux
versions, and most 64 bit version were only offered in Linux versions.
|
|
