IMPOSSIBLE!

[Tim Rue]

Quoting Alan Grimes <agrimes at speakeasy.net>:

> =P
> 
> Everything I have ever learned about the unix operating system has
> strongly indicated that non-trivial IPC acceptable for use in modern
> applications was completely impossible.
> 
> Even on versions of Unix based on *ghasp* modern kernels such as Minix,
> including state-of-the-art IPC mechanisms, Extreme lengths and pains
> have been taken to render said IPC utilities crippled and unusable.
> 
> There seems to be an article in the UN charter which prohibits the
> development of good operating systems such as L4 beyond the toy stage.
> 
> So, I beg to know, by what sourcery has the miracle of IPC been
> implemented on Linux?
> 
> Linux is still somewhere between Awful and gawdawful. 2/3rds of the
> programs I want to run will instantly crash my machine when I try to
> launch them. I've wanted to develop my own OS since 1994. I've pretty
> much given up, I have the design but I've been unable to establish a
> toolchain that will allow me to do what I need to do. =(

First off, a computer is a bit flipping machine, naturally constrained 
only by physics of the hardware, including physics of Integrated Circuits.
And then the human imposed programming constraints and we all know how 
unfair, not so honest, we can be.

The constraints you refer to are human based. And as such the problem 
you refer to is also human based.

There are several factors contributing to the changing acceptability of 
easier access and use of what is in essence the third primary user interface. 
The first two being command line or shell interface and the second being the 
general GUI concept interface. And like colors of paint or light, you remove 
access to just one of the primaries and you greatly limit what you allow 
others to do for themselves. 

The "why would anyone do this" is simple: 
Power over others, money, social status, control, ego, etc.

Perhaps the most influencial factor of change is the growing threat of 
software patents, where two of the factors considered in granting a software 
patent is that of being novel and inventive in the scope of those skilled in 
the art. In other words, not obvious to those skilled in the particular 
application field. Like a one click shopping method would have been obvious 
to any end user experiencing the hassel of online shopping, had only they had 
the tools (genuine honesty and openness about these three primary interfaces), 
to come up with it themselves in mass.

The historical analogy to the change is that of the move in mathmatics from 
the roman numeral system to the hindu-arabic decimal system that enabled the 
general population to perform math calculations beyond the complexity of what 
the roman numeral accountant elites were able to do. It took three hundred 
years to convert but the hindu-arabic decimal system is perhaps the most 
univerals language used today. The delay is due resistance by those with 
vested interest. Other examples of delays in change for the better include 
Galelio and the catholic church. But the change only happened in order to 
benefit the pursecutors, for Galelios exoneration in 1992 didn't help him 
during his life but help bring back a following that was dwindling.

D-bus is a good step in the right direction but is not without complexities 
that will soon become simply not so much needed for most systems such IPC 
is used on and by the user.

As memory/storage devices become larger in storage amount, faster in speed 
and smaller in size and cost, the use and need of multi-user systems 
decreases, or at least the software side does. Systems today can boot off 
a USB firefly size drive that has been set up as a single user system. 

The security overhead in dealing with multi-user systems is overhead not 
needed in single user systems.

The other factor in need of change is making the functionality of software 
packages accessible through such IPC mechanisms as D-bus, and in a user 
friendly manner. This is more likely to happen via open source software 
then it is with closed source proprietary software. Likely because open 
source licenses such as the GPL do not have the apparent UN article you 
mention, or at least its not ....directly.... enforceable against any 
programmer who decides to alter the code. Maybe we just need new rebels 
to take us to the next stepping stone in teh evolution of computing....

The Amiga OS (before it bacame a con artist label) had all three user 
interfaces available to the user. The third best mistakenly known as the 
AREXX port. And people did put "toolchains" together to accomplish things 
for themselves. But Amiga was apparently shot down by the UN article you 
mentioned and further deteriorated by the turning it over to con artist 
to obviously abuse its followers.

You mentioned interest in creating your own OS but lack the resources. 
Perhaps you'd be interested in contributing to OSS projects like AROS or 
Dragonfly BSD which share in the simpler IPC concepts inspired by the 
real Amiga. 

Or maybe what you are really looking for is the application of 
Abstraction Physics http://threeseas.net/abstraction_physics.html
in a manner that openly recognizes abstraction physics. Perhaps you 
might be interested in contributing to the tool based upon abstraction 
physics project, the Virtual interaction Configuration. 

Ultimately what you seem to be wanting is the ability to easily design 
and produce software for your personal use that doesn't crash. Software 
that may very well make use of functionality others have created within 
some program they created or contributed to. The ability to apply the 
concept of reuse of functionality with option of dealing with it at the 
code level or NOT. "NOT" being the use of the functionality thru the 
application it is currently found in, via IPC third user interface 
to that functionality.   

Tim Rue (3seasdbus)
----
To protect his social status a roman numeral accountant 
would have promoted: "Only a fool would think nothing 
can have value." (re: zero place holder, decimal system)