A few things

[Walt Pawley]

On 9/7/04 2:34 PM -0500, Clinton MacDonald wrote on Re: A few things

>That's right. What is a crying shame is that compiled binaries (or
>"packages") are often not compatible between different Linux
>distributions on the same platform. If Linux is going to become a
>"desktop OS," it will have to overcome these kinds of issues -- issues
>that make Linux seem just as "proprietary" as closed source software.

Okay, so you dragged me, kicking my feet and scraping my fingernails to the
quick trying to stay out of it ... sigh.

We've long had the ability to avoid proprietary "binaries" but no one
(well, other than a few really strange dudes, some of whom I occasionally
resemble) has taken such technology seriously. The only thing that makes
Linux "open" is that the source code is available. For almost all of the
six or so billion of us, this not directly important as we'd have little
chance of doing much of anything cogent with it in the first place (if
./configure; make; sudo make install doesn't work, we're toast).

This is not all bad. As

On 9/7/04 5:55 PM -0400, Ed Sutherland wrote on Re: A few things

>... One of the
>major hindrances to Linux becoming a desktop OS (rather than mostly a server
>or terminal) is the incompatibility of packaging schemes.

Perhaps the biggest problem with these schemes is that they're largely
opaque, at least to most of us. But a major issue of them all is that they
are designed to support much more tightly defined environments than we'd
like to throw at them. One can conceive of a source code level
installation system that would be highly resilient, needing little more
than a shove (the ubiquitous double-click, say) to do its deeds whether
installing or uninstalling. FWIW, Apple makes a stab a the last part if you
ignore all the extra mumbo-jumbo about legalities. Fink sort of makes a
stab at the first part, but not in general - just for Mac OS X, I believe.

What's needed is to QUIT making new system level stuff for a while.
Blasphemous, I know. But it's necessary to settle things down long enough
for a foundation design to take root. I, for one, don't expect any such
sanity to erupt.

In the end, it may not matter all that much. As computers become ever more
capable of dealing with masses of data, we may quit typing on keyboards
and, instead, carry on conversations with our silicon assistants WHO will
deal with these and other issues themselves.

-- 
Walter M. Pawley <walt@wump.org>
Wump Research & Company
676 River Bend Road, Roseburg, OR 97470
         541-672-8975