Created 30. March 2004.
A speculative look into the future
Necessity makes strange bedfellows. As an example, let us go to the
magical time of late 1980s. At that time two companies fought fiercely
over the game console market. Those companies were Sega and Nintendo.
The atmosphere could be called openly hostile. Yet, fast forward some
ten plus years and we find that these two companies are very close
corporate partners.
This simple example hopefully demonstrates that seemingly impossible
things happen, if there is a sound business reason for it.
This phenomenon is not limited to business, but also applies to other
aspects of life, such as politics.
In this article I have tried to think of events that are currently
very unprobable, but may in fact happen some day. Read through them
and think about them yourself. Are the ideas majorly stupid? Could
they happen? When? What would be the implications of these events?
What other unlikely events can you think of?
First patch to Linux kernel from Microsoft
Very strictly speaking this might already have happened. Some
Microsoft employee might hack on the kernel somewhat on his free time.
But in this case I mean an officially sanctioned patch developed by a
MS employee on company time. While kernel patches are quite unlikely,
patches to user space programs might be possible.
A major desktop/laptop manufacturer sells only Linux machines
Since none of the current major sellers could afford not to sell
Windows machines (even if they wanted to), this calls for some new
company. This can only happen if some current or future small Linux
laptop maker grows enough to challenge the Dells and HPs. This would
require Linux's global market share to be quite large, something like
30% or more.
Software patents are abolished
This one requires either a big swing in the political climate or
hundreds of millions of dollars worth of losses to big companies
(maybe due to junk patent vampirism by litigious companies). Another
possibility is a humongous patent battle between two huge parties.
Since it is widely known that big software houses have lots of patents
on obvious stuff that pretty much everyone uses, a small lawsuit could
expand to a huge inferno resulting in lots of smoking ruins and filthy
rich lawyers. I have to admit that a no-holds-barred patent smackdown
between IBM and Microsoft is an fascinating idea in some morbid way.
X gets replaced
The heading above is just about as old as X itself. The main arguments
against changing X is that X works and the porting costs would be
enormous. The first one pretty much still holds its own, but the
second one is not as important anymore. Some years ago a lot of X
programs used either Xlib or Athena widgets. Porting these kinds of
programs to non-X platforms takes a lot of work. Most programs today
are coded with modern, multiplatform toolkits, like GTK+ or QT. This gives a migration path
away from X. These programs are just a recompile away from porting,
since they don't have direct dependencies on X libraries. However,
given the evolutionary nature of Unix, the X protocol most likely
remains for the foreseeable future, but everything around it gets
redesigned.
A 3D graphics card with fully open source drivers is introduced
The current state of 3D accelerated graphics cards on Linux is very
non-optimal NVidia's video drivers
are not open source and Ati's
drivers, well, they are not nonexistant. There exists a market for
a open source graphics card. An alternative to providing open source
drivers is to fully document the card so the open source people can
write their own drivers. While there are business (meaning secrecy)
reasons in not revealing how hardware works, there are also
advantages. Firstly, the said cards would most likely become popular
among Linux people, even if they weren't the fastest ones around.
Secondly you get a huge group of expert beta testers who can and want
to track hard to find bugs and maybe even create fixes for them.
A new programming language gains popularity
C and C++ are the cornerstones of the computer industry.
However they have several drawbacks considering modern systems. If a
new language appears that hits one of these sweet spots, it has a
chance to become popular. What could these features be? One irritating
deficiency of C/C++ is the amount of rope the preprocessor gives you.
For example there is no simple way of knowing what portions of source
code are function calls without parsing the entire file. This makes
integrated IDEs very hard to build. Note that I'm lining out languages
based on virtual machines, like Java. They have their places, but it's
not this one.
FSF relicences all its libraried to GPL
Currently many basic libraries that make up Linux are licenced under
the LGPL. This
means that closed source programs can link to them freely. Now suppose
the FSF decides that all future releases are under the GPL. This would cause
quite a ruckus among Linux developers. Some people would fork the old
LGPL versions and develop them further. Others would perhaps try to
port the corresponding libraries from the BSDs. Yet other people would
start a new be all, end all project to create all new libraries
from scratch. All these would most likely be binary incompatible,
paving the way for a mountain of lost hair.
Afterword
The above text speaks only of Linux. This was a choice dictated by
clarity. In most places "Linux" could be replaced with any free
operating system, such as NetBSD
or Syllable.
(C) 2004 by Mr Shrap. All rights reserved.
Back to front page.