reasons for new Winsock TCP/IP implementation, and Winsock with TNC wit

[mikebw@bilow.bilow.uu.ids.net (Mike Bilow)]

Gary Crum wrote in a message to Mike Bilow:

>Given this situation, the last thing we need is to reimplement a parallel
>and probably conflicting Winsock stack of our own in order to get access
>to AX.25.

 GC> Here are a couple reasons to build a new Winsock
 GC> implementation based on NOS:

 GC> 1. Trumpet Winsock is shareware, not free.  Does anyone know
 GC> of an officially free Winsock?

Microsoft distributes a free Winsock stack with Windows for Workgroups that is
going to be in Windows 95 when and if it ever comes out.  OS/2 includes a
TCP/IP stack and a shim to make it accessible using Winsock from within Windows
on OS/2.  As with Unix, TCP/IP soon is going to be regarded as part of the
operating system.

 GC> 2. Such a Winsock IP implementation derived from the KA9Q
 GC> NOS implementation   would naturally support multiple
 GC> network interfaces, and IP routing.  Does   anyone know of a
 GC> Winsock module that does these things?

Since the OS/2 Winsock implementation is really just a shim to access the
native OS/2 TCP/IP stack, multiple interface support is no problem provided
that LAPS or some similar OS/2 facility is available.

 GC> P.S.  Here's a third and perhaps most important reason:
 GC> Freely available   source code to a Winsock-compatible
 GC> TCP/IP implementation will foster   interesting development.
 GC>  This be quite in line with a primary FCC-designated
 GC> purpose of amateur radio frequency allocations --
 GC> communications technology   development.  This in FCC Part
 GC> 97.1.  Go, FCC, the most improved government   department
 GC> this last year! -- except for frequency auctioning perhaps.

I would prefer to use Linux or FreeBSD for that kind of research.

-- Mike