Re: evolution of nos
- To: firstname.lastname@example.org
- Subject: Re: evolution of nos
- From: email@example.com
- Date: Thu, 16 Mar 95 10:41:59
- Reply-to: firstname.lastname@example.org
> This may appear to be a rather simplistic, perhaps idealistic and
> obviously ignorant observation from a non-programmer... for the
> technically adapt, please go easy on the flames. My attempt is to
> generate some useful discussion. Is there any reason
> why we couldn't use existing winsockets or in the <future??> the
I dont think its an ognorant observation at all. There has been alot of work
done in this are, specifically for Linux, OS/2 and perhaps even Windows.
The problem, simplistically is.
As it is true that OS/2 and Linux and WinNT etc come with a (or for a
nominal fee) their own stack that is built around the underlying operating
environment, these are designed to support an ethernet interface primarily
and perhaps a slip interface. What they do not support is an AX.25 or KISS
interface. These interfaces belong buried within the stack code itself as
that is where the stack talks to the interface. In the case of Linux, the
source (I believe) is available for the entire OS, including the stack. In
this case on simply adds the code for the new interface and connects it to
the existing stack logic. In the case of OS/2 there is a programatic method
to connect a new interface with the stack, however it is not public domain
and is subject to change release to release. WinSock has precisely the same
problem and is more proprietary, secret and subject to change. Last I heard
that was precisely what Johan was working on. If not WinSock, a functional
equivalent that would have the proper code for AX.25 and KISS.
Whenever you deal with a real live software company that exists from the
profits of its software, you are not apt to be given access to their source.
FSF and GPL aside, most software vendors and ISV do not publish their
source. What makes matters worse is often this information is available via
executing a non-disclosure agreement with the vendor. In doing this the
individual may gain knowledge of how to build the liinkage to the base stack
but can not publish the source on how to do it, since it is based
priviledged knowledge. If other code they use is covered under a GPL then
they 'must' publish their source, which at this point they are legally bound
to keep secret. Consequently, in order to 'properly' handle this situation
they must write everything from scratch and not publish their code. So, I
believe why you are not seeing alot of this is its a long process and there
are few people willing to solve that 'problem'.