[3307] in linux-net channel archive
Re: Binary Driver Issues
daemon@ATHENA.MIT.EDU (Dennis)
Tue Jun 18 19:09:37 1996
Date: Tue, 18 Jun 1996 15:12:14 -0400
To: dledford@dialnet.net
From: dennis@etinc.com (Dennis)
Cc: linux-net@vger.rutgers.edu
>
>On 18-Jun-96 Mike Kilburn wrote:
>>>
>>> I would assume that most linux developers have jobs which they do work that
>>> they dont give
>>> away. Do they have double standards if they use linux? Are you saying that
>>> no commercial
>>> company should be allowed to use linux as a server because they are
>>> benefitting for free?
>>> Are you saying that all Linux-based ISPs should be banned from using linux
>>> because they're
>>> making money from it? Or that anyone that writes a script for Linux for one
>>> of their customers
>>> machines and does not submit it is stealing?
>>
>>No. You misunderstand. You cant enhance the software and sell it without
>>giving back source. Using it is no problem. And I am talking about the
kernel.
>>If you want to
>>interface in user mode thru libc then binary only is OK. Linus has allowed
>>this loophole in the kernel GPL but I dont agree with it.
>
>Actually, this is somewhat incorrect. Take for instance the USS sound driver.
>Linux includes this driver in the regular distribution, but it is a pared
>downed driver without full functionality. In order to get the full
>functionality, a person must buy the commercial version from the author.
>Whether this commercial version is distributed as binary only I'm not sure
>about. ET could do the same thing. Allow a driver for their card in linux.
>Don't include the bandwidth management features and such with that driver.
>Allow the basic Frame Relay driver in 2.0.0 to be the higher layer to their
>card. Then, if someone wanted the full functionality of their current driver
>they could buy the enhancement module in binary only form from ET. This module
>could be made in such a way that it interfaces with the card driver in a static
>fasion so that kernel upgrades and version changes would not effect it. I
>would see this as a perfectly acceptable solution to the problem. It would
>protect the intellectual property of ET, while at the same time providing a
>basic driver for linux, and giving source for the basic driver so that it could
>grow and change with changes in the kernel structures.
Yeah, we'll just re-design our product so you can have a useless source driver
for the distribution. Great idea!
db
