Posted: Thu Mar 30, 2000 4:55 am Post subject: [Asterisk] Re: H.323
IMHO, the discussion about licensing conflicts misses the point somewhat.
Nobody wants to incorporate OpenH323 into Asterisk or vice versa.
Linux itself, for instance, is licensed under the GPL. However, nothing
prevents my program to talk to Linux by way of calling its published
entry points (this includes binary-only kernel modules, which are
dynamically loaded into the kernel just like Asterisk's subprograms).
Mr. GPL (Richard Stallman) has, AFAIK, never complained about that.
Thus, if I write a program which happens to compile to a .so file which
asterisk can load and which happens to be dynamically loadable into
Asterisk, nobody in their right mind could possibly complain about this.
It's my own program, published under my own license; I can and will do
what I want with it, and -- since it doesn't have to be part of either
the Asterisk or the H.323 efforts and the license will allow it -- so
can everybody else.
Does anybody want to help?
--
Matthias Urlichs | noris network GmbH | smurf@noris.de | ICQ: 20193661
The quote was selected randomly. Really. | http://www.noris.de/~smurf/
--
It's interesting to think that many quite
distinguished people have bodies similar to yours.
Posted: Thu Mar 30, 2000 5:20 am Post subject: [Asterisk] Re: H.323
Quote:
Thus, if I write a program which happens to compile to a .so file which
asterisk can load and which happens to be dynamically loadable into
Asterisk, nobody in their right mind could possibly complain about this.
It's my own program, published under my own license; I can and will do
what I want with it, and -- since it doesn't have to be part of either
the Asterisk or the H.323 efforts and the license will allow it -- so
can everybody else.
On the contrary. First of all, you would have to use Asterisk to build
such a module (for headers, for example). Supposing you didn't and that
you built everything from scratch, your .so would be useless without
Asterisk and thus could arguably be considered to be a part of it -- this
isn't as solid a foot as I would like to have, but if push came to shove,
I would probably bring a case to see what would happen.
You definitely could not distribute your module with Asterisk since that
*would* be linking (dynamic or static) without distributing everything
under GPL.
Posted: Thu Mar 30, 2000 6:29 am Post subject: [Asterisk] Re: H.323
Hi,
Mark Spencer:
Quote:
On the contrary. First of all, you would have to use Asterisk to build
such a module (for headers, for example). Supposing you didn't and that
The same happens when I build a module for the Linux kernel.
There are lots of binary-only kernel modules out there. Nobody likes
them much, true, but nobody's suing the makers of these modules either,
precisely because Linus has said that he considers them OK if they only
use the public interface.
You could do the same.
IMHO there are three choices here.
- No H.323 at all.
- Write your own.
- Write a program which uses both, and distribute it separately.
The first is nonsense, frankly. So is the second, simply because nobody
has the time to do it. I wouldn't do it if I had all the time in the
world because an open-source is available, and that's that.
Which leaves us with the third alternative.
Quote:
You definitely could not distribute your module with Asterisk since that
*would* be linking (dynamic or static) without distributing everything
under GPL.
I don't _want_ to do that. I know that the licenses are incompatible.
Thus, the question is whether you want a program that's usable by many
people, or whether the Purity Of The Sacred GPL As Interpreted By RMS is
more important.
For me, that question is a no-brainer. If I'm not changing the source
code of programs A or B in _any_ way, my publishing of a source code C
which happens to use them both as-is (which hopefully is possible...)
isn't a question of licensing -- it's a question of free speech.
--
Matthias Urlichs | noris network GmbH | smurf@noris.de | ICQ: 20193661
The quote was selected randomly. Really. | http://www.noris.de/~smurf/
--
The existence of god implies a violation of causality.
On Thu, Mar 30, 2000 at 08:29:57AM +0200, Matthias Urlichs wrote:
Quote:
There are lots of binary-only kernel modules out there. Nobody likes
them much, true, but nobody's suing the makers of these modules
The only reason binary-only modules are acceptable is that Linus said
they were, not because the GPL allows them.
--d6Gm4EdcadzBjdND
Content-Type: application/pgp-signature
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v0.9.9 (GNU/Linux)
Comment: For info see http://www.gnupg.org
Posted: Thu Mar 30, 2000 2:53 pm Post subject: [Asterisk] Re: H.323
Quote:
Thus, if I write a program which happens to compile to a .so file which
asterisk can load and which happens to be dynamically loadable into
Asterisk, nobody in their right mind could possibly complain about this.
It's my own program, published under my own license; I can and will do
what I want with it, and -- since it doesn't have to be part of either
the Asterisk or the H.323 efforts and the license will allow it -- so
can everybody else.
Um, no. Sorry. GPL makes a distinction between stand-alone code
and coded that is linked into another program, making one larger
program. Linux modules is a gray area, blessed by Linus so no
one yells loudly. However, libraries are distinctly covered -
hence the existance of the LGPL.
Now, if the code you want to GPL and link to some MPL code is
*100%* yours, you can name an exception in your licence that
allows it to be GPL in all cases except where linked to that PL
library, in which case you allow MPL for the purposes of that
combination only. It's your code, you can allow that.
The problem arises when you have LOTS of contributors to a
project that have contributed code under one license... to
legally allow an exception would require getting permission from
all contributors to the project. Wow. That can be almost
impossible if the project has been around awhile (like OpenH323).
Greg
/********************************************************************
Greg Herlein Quicknet Technologies, Inc.
Director 415-512-1306
gherlein@quicknet.nethttp://www.quicknet.net
*********************************************************************/
Posted: Thu Mar 30, 2000 4:47 pm Post subject: [Asterisk] Re: H.323
On Thu, Mar 30, 2000 at 12:30:06AM -0800, Charles Duffy wrote:
Quote:
On Thu, Mar 30, 2000 at 08:29:57AM +0200, Matthias Urlichs wrote:
> There are lots of binary-only kernel modules out there. Nobody likes
> them much, true, but nobody's suing the makers of these modules
The only reason binary-only modules are acceptable is that Linus said
they were, not because the GPL allows them.
This is something I've wondered for a while, How does Linus have the
right to make that exception to the GPL when Linux has hundreds of
contributers?
Posted: Thu Mar 30, 2000 5:52 pm Post subject: [Asterisk] Re: H.323
Quote:
This is something I've wondered for a while, How does Linus have the
right to make that exception to the GPL when Linux has hundreds of
contributers?
IIRC, he made the decision very early on in the life of Linux, before there were many (any?) other contributors - he probably made the decision around the time the license changed to GPL (the first versions of linux weren't GPL'ed - I think they were just available for anyone to look at/use etc.).
Posted: Fri Mar 31, 2000 9:28 am Post subject: [Asterisk] Re: H.323
I think the simplest solution to the "licensing problem" is to create an open point-to-point protocol/interface. Just like TCP/IP it would become available to many developers who can link there applications with Asterisk or OpenH323.
I suppose such protocols allready exist. In the telecommunications world PBX's can communicate with each other (ITU standards like SS7) and maybe the same protocols could be used to glue OpenH323 and Asterisk. We probably don't need all the fancy features, but it can be a place to start.
Unfortunately I don't know much about the availability of such protocol specifications.
The use of a protocol means you perhaps don't have to run OpenH323 and Asterisk on the same system. This way you can create an open, flexible and distributed PBX.
Posted: Fri Mar 31, 2000 12:20 pm Post subject: [Asterisk] Re: H.323
Hi,
I.H. van der Molen:
Quote:
The use of a protocol means you perhaps don't have to run OpenH323 and
Asterisk on the same system. This way you can create an open, flexible
and distributed PBX.
It makes sense to put it all into one program because transmitting data
on a network, or between two unrelated processes, requires additional
buffering. Buffering increases latency. Anything that increases latency
is a Bad Thing.
Besides, there already is a simple transmission protocol (IAX).
Why reinvent the wheel?
NB: An implementation of SS7 for Asterisk would be a cool thing to have,
but it would be useful for talking to a T1 card in your PC, not across
Ethernet. And it would need Telco approval, which is not at all easy
(or cheap) to obtain.
--
Matthias Urlichs | noris network GmbH | smurf@noris.de | ICQ: 20193661
The quote was selected randomly. Really. | http://www.noris.de/~smurf/
--
If it's Tuesday, this must be someone else's fortune.
Posted: Fri Mar 31, 2000 2:32 pm Post subject: [Asterisk] Re: H.323
Quote:
The same happens when I build a module for the Linux kernel.
Long ago Linus made a shaky, but well known exception permitting binary
modules for the Linux kernel that used the published API's. Please note
that this is not something allowed in the GPL, but something which Linus
in particular (to the chagrin of RMS) allows.
Quote:
There are lots of binary-only kernel modules out there. Nobody likes
them much, true, but nobody's suing the makers of these modules either,
precisely because Linus has said that he considers them OK if they only
use the public interface.
Yes, he did that, but that does, of course, does in theory make the Linux
kernel license not compatible with GPL, but of course everyone overlooks
it.
Quote:
You could do the same.
First, I do not see why I should be willing to make the same exception for
Asterisk that was made for binary modules in the Linux kernel, but feel
free to correct me.
And again, if I do, I will NOT be able to use other GPL'd code in my
program which has not made the same exception. It's the exact same
problem as the OpenH323 exception, and I cannot understand why there is so
much trouble understanding this point. The good and bad thing about the
GPL is you either take it entirely as it is, or you don't take it at all.
Quote:
Which leaves us with the third alternative.
Right, so somebody go code it :-P I've already said I'd be willing to let
the IAX code go out under another license (like GPL+MPL)
Quote:
Thus, the question is whether you want a program that's usable by many
people, or whether the Purity Of The Sacred GPL As Interpreted By RMS is
more important.
It's not merely a theoretical, but a very practical question.
Quote:
For me, that question is a no-brainer. If I'm not changing the source
code of programs A or B in _any_ way, my publishing of a source code C
which happens to use them both as-is (which hopefully is possible...)
isn't a question of licensing -- it's a question of free speech.
You would obviously have to make some code to make Asterisk use OpenH323.
Anyway, as I've said, there's no point keeping this licensing debate going
until there's some code to make it happen, and at this point, nobody has
stepped up to the plate to code either an H323<->IAX gateway, nor an
OpenH323 channel module. When there's code, we'll work it out one way or
another.
Posted: Fri Mar 31, 2000 4:22 pm Post subject: [Asterisk] Re: H.323
Hi,
Mark Spencer:
Quote:
Right, so somebody go code it :-P I've already said I'd be willing to let
the IAX code go out under another license (like GPL+MPL)
Good.
Quote:
Anyway, as I've said, there's no point keeping this licensing debate going
until there's some code to make it happen, and at this point, nobody has
stepped up to the plate to code either an H323<->IAX gateway, nor an
OpenH323 channel module. When there's code, we'll work it out one way or
another.
The idea to write a module seems more attractive. I've put this idea
somewhere on my "would be nice to code" list, but I don't know when I'll
have the necessary round tuits.
--
Matthias Urlichs | noris network GmbH | smurf@noris.de | ICQ: 20193661
The quote was selected randomly. Really. | http://www.noris.de/~smurf/
--
In a whiskey it's age, in a cigarette it's
taste and in a sports car it's impossible.
Posted: Fri Mar 31, 2000 4:29 pm Post subject: [Asterisk] Re: H.323
Quote:
The idea to write a module seems more attractive. I've put this idea
somewhere on my "would be nice to code" list, but I don't know when I'll
have the necessary round tuits.
I don't know enough to know which idea is more attractive, the
OpenH323/IAX bridge or OpenH323 as a module, becuase I don't know enough
about C++.
Posted: Fri Mar 31, 2000 5:00 pm Post subject: [Asterisk] Re: H.323
Hi,
Mark Spencer:
Quote:
I don't know enough to know which idea is more attractive, the
OpenH323/IAX bridge or OpenH323 as a module, becuase I don't know enough
about C++.
The "as a module" option is more attractive to me, because the bridge
would involve another network connection while a module would just pass
data buffers around within the Asterisk program.
Whether the code organization of Asterisk and H.323 allow this kind
of integration, especially since Asterisk is a multithreaded program,
remains to be seen.
--
Matthias Urlichs | noris network GmbH | smurf@noris.de | ICQ: 20193661
The quote was selected randomly. Really. | http://www.noris.de/~smurf/
--
Dr. Caligari's Come-Back:
A bad sector disk error occurs only after you've done several hours of
work without performing a backup.
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum