Re: [tlug] [OT] Say _no_ to the Microsoft Office format as an ISO standard

[Curt Sampson <cjs@example.com>]

On Tue, 24 Jul 2007, Stephen J. Turnbull wrote:

> I don't think the BSD stack had that much to do with spiking BitNET,
> ChaosNET, DECNet, and so on.

I'd say that they were all pretty much dead, anyway. Not being available
on microcomputers in a practical way by 1995 pretty much took you out of
the competition.

> I'm not even sure it really had that much to do with spiking NetBIOS
> and AppleTalk.

NetBIOS is a different layer; it was rapidly ported to run of TCP/IP,
and that pretty much became the default configuration, since everybody
wanted TCP/IP up and running to browse the web and collect mail from
their ISP, anyway.

As for AppleTalk: you tell me what Apple currently recommends you use
instead of it. :-) (Though that one took a while, I'll admit, mostly due
to missing features in the IP protocols.)

IPX failed as much technically as for any other reason: their
internetworking just wasn't suited to large internetworks; you can
stretch stuff designed for fast local LANs just so far.

> I think that the growth of the Internet was really very much about the
> "best current practice" approach that the IETF has always taken to
> standards.

That made it work, but I'll argue that it didn't make it popular. What
did was:

> But the availability of the BSD stack definitely made Microsoft's (and
> others') decision to support TCP/IP a no-brainer...

Think about it. You're a vendor of some sort of network system, with
your own semi-proprietary protocol (though Cisco supports it, of
course), it's 1996, and you're faced with updating your network stack
once again, and dealing with all those driver installation issues
too. Then you realize that since last November, 90% of all of the
new microcomputers that will be installed from now on will have a
well-supported TCP/IP stack built in. What are you going to do? You
wouldn't even need free code for the stack available, by that point; the
advantage of chopping out a whole lower slice of the protocol stack, and
one of the most painful parts to install, would be enough to make you do
whatever you could to get your stuff working, by hook or by crook, over
TCP/IP.

> > It's interesting to think that had that stack been GPL'd, Stallman might
> > still be distributing gcc via magtapes through the mail today.
>
> <shiver mode="uncontrollable">

Well, my point was, it certainly would have rather blunted the success
of the whole Gnu project. Basically, it was non-GPL'd software that gave
him the distribution and collaboration system he needed to have Gnu
software become the big success that it is.

It almost makes you wish the license forbid the distribution of Gnu
software using non-GPL'd code. :-)

cjs
--
Curt Sampson       <cjs@example.com>        +81 90 7737 2974
Mobile sites and software consulting: http://www.starling-software.com