The Register - There are still crypto reg's...

[Phillip H. Zakas]

I completely concur there's a feedback loop problem, but its Apple's fault I
think.  I remember when the first MACs came out you had to pay $5K just for
the privilege of programming for it.  What numbskulls!  The intel platforms
were the first to encourage development because bios ref. guides were cheap
and most could afford the $100 of a pascal, c or asm compiler.  Plus the
intel-platform hw (ibm, compaq, etc.) was really designed to handle
multitasking and simultaneous networking/communications.  Apple only
recently started to get the hint and improve the hardware.

As for the P2P stuff, I think you're right here too: it should be everywhere
all the time. Every linux distro, etc. should include it.  That's a great
way to hit critical mass.  But I do think that a really useful app will find
its way across the internet.  Remember ICQ before AOL bought it?  It was one
of the few things that hit big and hit fast (partly because AOL's IM service
wasn't available yet.)  Napster hit the same way.  Gnutella was never the
kind of thing the average non-programmer 19 year old cares about so it has,
and will continue to have, a small user base.

As for the open source comments, I do believe Linux is a truly powerful
force, but I am skeptical that most of the people 'peer reviewing' code
actually know what they're talking about.  Do they really know what
threading is?  I'm always surprised to find script kiddies borrowing code
frankenstein-style ('frankensteining code' in my own description) to piece
together really poor stuff that 'works' but not very well.  I'm less worried
about malicious code in linux than I worry about getting code that works.
In contrast openbsd has a more targeted programmer base and seems to be
developing in a more stable fashion.  Anyway I'll continue to work with
linux because i recognize the market drives technology but i do wish we'd
find better programmers for linux.
phillip




-----Original Message-----
From: owner-cypherpunks at Algebra.COM
[mailto:owner-cypherpunks at Algebra.COM]On Behalf Of Ray Dillinger
Sent: Thursday, February 01, 2001 7:08 PM
To: Phillip H. Zakas
Cc: cypherpunks at einstein.ssz.com
Subject: Re: RE: The Register - There are still crypto reg's...





On Thu, 1 Feb 2001, Phillip H. Zakas wrote:


>Apple is a pretty computer and everything, but there's so much to be
desired
>from their os programmers (not saying microsoft is the best, just saying
the
>mac os really stinks.)

There's kind of a feedback loop here; Windows achieved critical
mass and Apple didn't.  Hence, MS can afford to hire lots more
engineers to work on their CRAPPY os than Apple can hire to work
on their mediocre os.  So, after a while, Windows got better than
the MacOS.

I remember how hard Win3.11 sucked compared to MacOS 5.  But once
true multitasking was under WinNT3.51, it was about Neck-and-neck
for quality with MacOS 7.  And like Mr. Zakas, I'm pretty convinced
that even though MacOS 10 has true multitasking, it has definitely
fallen behind WinNT 4.

ObCryptoStuff; this same kind of "critical mass" phenomenon
affects lots and lots of the stuff we're looking at.  Freenet,
Mojo Nation, Napster, Gnutella, Digital Cash, etc, even PGP and
GPG - All have a value that depends directly on how many other
people are using them. If they don't grow beyond a certain
threshold size, they remain less useful than the disk space they
occupy.  But where is the threshold and when is it passed?

And how do you get there?

If P2P agents were distributed with a couple of the major linux
distributions (say Red Hat and SuSE) they'd probably achieve
critical mass fairly fast (especially if they were set up and
configured during install, the way Apache increasingly is).
And if they were opensource, they'd probably get into ALL linux
distributions within a few years.

But so far most of the people doing P2P are trying to make a buck
off of it, so with the exception of gnutella, there aren't open-
source agents.

				Bear