🚀 go-pugleaf

nevin@CS.Arizona.EDU (Nevin ":-]" Liber) wrote:
]
] First off, Linux is written by people who care about what a *Unix* system
] should be.  You can do far, far better from an operating system point of
] view if you don't constrain yourself to staying compatible with Unix.

Please provide evidence for this statement.

] Companies like Apple, on the other hand, work hard on providing new
] technologies.  Things like QuickTime, PlainTalk, etc.  I've never seen
] this kind of stuff come out of the GNU community.  I'd be far more
] impressed and far more embracing if I saw them pushing the envelope
] instead of small refinements; innovating instead of just evolving.

QuickTime is from SGI.
PlainTalk is fluff.

What about GhostScript and GNUStep?  What about the OpenGL clone?

(Look ma, I'm defending the GNU community!)


] My Unix environment hasn't fundamentally improved (or heck, even
] changed in terms of user experience) in the last 5 years.  My
] Mac environment has, in both terms of user experience and
] development tools.  You might not like the underlying OS; then
] again, except for developers, most people don't see it.

You are mistaken about the UNIX environment. 5 years ago it was
early 1990.  Most UNIX kernels did not have module loading.  They
also did not have shared libraries.  They did not have multicast
support in networking.  They did not have the cability of
multihoming using ethernet aliases.  They didn't come with X.

UNIX is very dynamic, contrary to your opinions.

] You guys complain that cooperative multitasking is *bad*, yet
] cooperative networking like TCP (eg: congestion control on the
] client side) is *good*?

First of all, TCP is not "cooperative networking", unless you are
trying to use "cooperative" in two different meanings.  Just because
there is an agreed upon protocol does not make it voluntarily
preeemptive (the meaning of "cooperative" relative to multitasking).

Second, congestion control mechanisms like "source quench" which
happen to occur on the client side are relatively new.  They are
also forced upon us by common carriers that do bandwidth overcommit
-- basically sell more bandwidth than they have in the hopes that
it won't necessarily be used.  And this is neither a general good,
nor is it a desirable state of affairs for anyone by the common
carriers, and only then because they isn't serious competition
in that market (hint: wire/optical fiber ownership).

] Or, more likely, wait for companies like Apple to try something new
] out, and if people like it, then maybe someone will port it over
] to Linux.

Like Apple invented preemptive multitasking and memory protection
for Copeland?

Or like they invented GUI interfaces?


                                        Terry Lambert
                                        terry@cs.weber.edu
---
Any opinions in this posting are my own and not those of my present
or previous employers.