RE: [oc] Re: ask slashdot?

["Stackelhouse, Scott (FL51)" <scott.stackelhouse@honeywell.com>]

I guess I don't understand why someone would need to expose code they aren't
allowed to, just because a portion of their design is GPL'd.  Yes, if the
*interface* is something that is secret then you have some trouble, but
nothing that couldn't be fixed by another level of abstraction--an interface
to the interface.   And yes, timing issues and space might preclude this
interface to an interface.  But so what.  That is a limited case to begin
with, and people more creative than I am will probably find a way to deal
with this.

Tivo (www.tivo.com) uses the Linux kernel and some source, but they also use
a lot of proprietary source.  It is all one combined mesh.  They don't make
their proprietary source available, but the GPL source they use is
available.  No conflict.  This thing is pretty neat by the way (Tivo), and
no I am not affiliated with them.

We need a license before we put code in a publicly accessible cvs.  Perhaps
an initial license to restrict all public use until we come up with a final
license is the way to go.

opencores.org has put out a call for designers around the world to chip in
(sorry about the pun) and help out if they are interested.  perhaps a public
call to legal entities for pro bono work regarding this issue would not be a
bad idea.

--Scott


-----Original Message-----
From: Victor the Cleaner [mailto:jonathan@canuck.com]
Sent: Wednesday, March 08, 2000 9:40 PM
To: cores@opencores.org
Subject: Re: [oc] Re: ask slashdot?


> > Personally I think its time to contact Richard M. Stallman at
www.gnu.org
> > and get him to write a new GNU license for HDL.  This would solve all
the
> > problems.
> 
> I am too sure about this... I am not saying this to offend anyone, but I
> personally think that Stallman is a bit over the edge.

I agree.  With all due respect to rms, this is not the time to call 
rent-a-zealot.

I've got to wade in here, because this is a fundamental issue that will
determine for many real-world designers (like me) whether the open-core
movement represents a Good Thing in the short term (long term - see below).


My basic feelings are this:

GPL makes it impossible for me to use an open-source core in my design if 
there's also a proprietary core in there that I've licensed from someone 
else.  That's going to make it much harder for the open core people to 
achieve market penetration, acceptance, credibility, etc., because until 
there are comprehensive libraries built and available, designers are going 
to be relying on (commercial) third-party cores for complex functions that 
we don't want to write ourselves.  This is a chicken-and-egg problem.

I mean, am I supposed to shitcan my entire design for a PCI board because 
I don't have the (market window) time and/or experience to write the bus 
interface myself, I can't use a commercial core because the GPL would force 
me to make public code I'm not entitled to do that with, and a VERIFIED 
open-source core is two years away and completely at the mercy of 
temperamental volunteers who have overcommitted themselves?

Hell, no.  Open cores lose. 

Trust me on this, though:  This question will not be "answered" here or
now.  Whichever side you choose will not satisfy everyone.  If you go GPL,
designers like me who favour the BSD license model will ignore you and go
start another group that satisfies our need to *choose* what, if, and when
we make source public.  If you go BSD, you will alienate the GPL "purists"
who are more obsessed with some sort of open-source philosophy than they
are with the day-to-day practicalities and choices faced by designers who
are not lifetime residents of universities and funded research institutions.

In other words, expect to be Balkanized just like the open-source Unix
movement.  But my perception of that world is that BSD has fewer camps, and
in those are found software of higher quality, where Linux is completely
out of control, with (at last report) 130+ distributions and counting.

Jonathan