Alexandre Oliva wrote:
Imagine if the reference TCP implementation had been GPL'd and no
commercial systems used it because of the restrictive license. We'd
still be struggling to make any two different systems communicate
today.
It's indeed difficult to implement code to follow specifications. Why
bother with Free Open Standards, let's just copy the code, right?
Yes, that's the point and value of licenses that permit it. Otherwise
there would be no reason to even discuss it. Everyone benefits. And
it's fairly clear that it happened that way - for example a few years
back there was a security fix that was necessary at the same time in
just about every device/OS that ran SNMP (Linux, most unix's, windows,
embedded systems in routers, etc.). Was it a coincidence that everyone
implemented the same bug?
Nevermind that the kernel Linux wouldn't take code under the original
BSD license, out of license incompatibility, and their TCP/IP stacks
could (and still can) interoperate.
And it took years of pain to shake the bugs out. Were you using Linux
seriously in the first few years of this crazy exercise? I had serious
data loss to several versions of NFS alone. How about that 'ping of
death' with an oversized packet?
So there... That implementation made life easier for all those OS
vendors who didn't want to respect their customers' freedoms, but
didn't help much other communities who did.
The *bsd implementation made life easier for everyone who had enough
sense to use it. Starting from scratch was an academic exercise that
put everyone involved through hell - and still - the best it can do is
exhibit standard behavior.
--
Les Mikesell
lesmikeseell@xxxxxxxxx
--
fedora-list mailing list
fedora-list@xxxxxxxxxx
To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
