Quote:> An X system is devoid of value if it is merely a protocol sitting in
> books on the shelf; it must be expressed as implementation in order to
> be of any use. Calling X a protocol may be correct, but ignoring the
> software is rather silly.
Well, i certainly didn't mean to be silly. Rather i was trying to express
my confusion about everyone's criticism. It would seem that if there is a
problem with Xlib, or a particular widget set, then the criticism should be
leveled at that offending party. Instead what i see is variations on "X
sucks". When there is a specific criticism at all, it is along the lines of
"X sucks because it doesn't do foo." And all i can think is gee, X
shouldn't do foo, but if you need to, write a library that allows you to do
it.
Quote:> All known implementations (and I'd be entertained to hear of any
> counterexamples) are based at some point on MIT and/or X Consortium
> code.
Ok. Again i am confused. Are you talking about the server or Xlib, or
both? But let's suppose that in my copious free time i want to reimplement
something from scratch. What? Is the problem with Xlib? Or is the problem
with the server? What should be reimplemented, and why? By "why" i am
asking what are the specific problems with the existing implementations, not
why should it be done.
Quote:> The point is that it is simply *too large* for a 386 system with 8MB of
> RAM to happily run along with any interestingly complex applications.
> And that is is the sort of "throwaway system of five years ago" that
> one sees.
I don't think (nor do you, if i've interpreted your web page correctly) that
this is a solvable problem. If you want all the features, you need a
certain minimum level of hardware. Fortunately that hardware is fairly
cheap these days.
[The link to the good enough is best essay in your unix01.html page is
broken, BTW]
--
nathan wagner "People with neckwear always have money."
-- looking for a *NIX sysadmin job within driving distance of Madison WI
resume at http://granicus.if.org/~nw/resume.html