From: Andrew C Aitchison <hidden> Date: 2004-01-31 13:12:52
On Sat, 31 Jan 2004, Sven Luther wrote:
On Sat, Jan 31, 2004 at 09:10:22AM +0000, Andrew C Aitchison wrote:
quoted
For several years the mga fb kernel driver has supported dual head and/or
dvi on cards which aren't supported by the XFree86 driver (unless you
use the mga_hal). I've wanted to use kernel code to add this support to
XFree86, but been put off by the licence problem.
And, have you asked the mgafb driver author about this ?
You can hardly complain about lack of back traffic if you didn't ask him
about it, and if you did, it would be interesting to this discussion to
know what the problems where.
"The Author" ?
This is open source code; there may be 27 authors of the relevant file.
In XFree86 code I wouldn't know how to find the author of a file without
looking at that file. My {limited ,mis}understanding of clean room coding
makes me wary of reading any source unless I know that its licence will
allow me to do what I wish.
OK. So I've probably been paranoid and lazy, but if the fbdev licence
had been compatible with the XFree86 one, I would have done the work.
As it is the bar was raised high enough to stop me.
quoted
So, for one developer at least, the reason there has been no traffic
from fbdev to XFree86 is *directly* because of the licence issue.
Yeah, but again, was it so because of a definite will on the fbdev
authors part, or because you didn't ask him ?
Isn't the aim of open source licences is to allow people to use the code
without tracking down the author and obtaining permission ?
I can do that with closed source.
--
Andrew C. Aitchison Cambridge
A.C.Aitchison@ntlworld.com
-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn
From: Benjamin Herrenschmidt <benh@kernel.crashing.org> Date: 2004-01-31 22:26:55
"The Author" ?
This is open source code; there may be 27 authors of the relevant file.
In XFree86 code I wouldn't know how to find the author of a file without
looking at that file. My {limited ,mis}understanding of clean room coding
makes me wary of reading any source unless I know that its licence will
allow me to do what I wish.
There is at least a maintainer and usually one person who did the
main work, and this (or these) name(s) can usually be found in the
file's head comment. Also, fbdev authors/maintainers usually hang
around the linux fbdev mailing list. You could have tried to
contact them at least ;)
Ben.
-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn
From: Ryan Underwood <hidden> Date: 2004-02-01 22:33:37
On Sat, Jan 31, 2004 at 01:06:23PM +0000, Andrew C Aitchison wrote:
quoted
And, have you asked the mgafb driver author about this ?
You can hardly complain about lack of back traffic if you didn't ask him
about it, and if you did, it would be interesting to this discussion to
know what the problems where.
"The Author" ?
This is open source code; there may be 27 authors of the relevant file.
In XFree86 code I wouldn't know how to find the author of a file without
looking at that file. My {limited ,mis}understanding of clean room coding
makes me wary of reading any source unless I know that its licence will
allow me to do what I wish.
The only authors that legally matter are the ones listed in the
copyright notices. There are three copyright notices in the matroxfb
stuff: one is Petr, another is Gerd Knorr and the other is Matrox. I'm
sure Petr has a pretty clear idea what he wrote, and I doubt Gerd would
get an attitude with you either over anything he did. Since you haven't
even inquired about it, you don't have much to complain about as I see
it.
OK. So I've probably been paranoid and lazy, but if the fbdev licence
had been compatible with the XFree86 one, I would have done the work.
As it is the bar was raised high enough to stop me.
Or maybe the license incompatibility was simply a convenient way to cop
out of doing some work?
quoted
quoted
So, for one developer at least, the reason there has been no traffic
from fbdev to XFree86 is *directly* because of the licence issue.
You can't copy and paste code. You _can_ rewrite code. Hardware
interfaces, trivial routines, and problems for which there exist only one
or a few obvious ways of solving them are all examples of where
copyright does not apply. Porting any code from fbdev to a XFree86
driver *will* involve substantial rewriting. I know this because I'm
currently doing it for the mga driver.
quoted
Yeah, but again, was it so because of a definite will on the fbdev
authors part, or because you didn't ask him ?
Isn't the aim of open source licences is to allow people to use the code
without tracking down the author and obtaining permission ?
No. The aim of open source licenses is to allow people to use the code
_under the terms the author chose_ without tracking down the author and
obtaining extra permissions.
I can do that with closed source.
I don't see how you have greater freedom with closed source licenses,
but feel free to elaborate.
--
Ryan Underwood, [off-list ref]
From: Sven Luther <hidden> Date: 2004-02-02 11:41:50
On Sat, Jan 31, 2004 at 01:06:23PM +0000, Andrew C Aitchison wrote:
On Sat, 31 Jan 2004, Sven Luther wrote:
quoted
On Sat, Jan 31, 2004 at 09:10:22AM +0000, Andrew C Aitchison wrote:
quoted
For several years the mga fb kernel driver has supported dual head and/or
dvi on cards which aren't supported by the XFree86 driver (unless you
use the mga_hal). I've wanted to use kernel code to add this support to
XFree86, but been put off by the licence problem.
And, have you asked the mgafb driver author about this ?
You can hardly complain about lack of back traffic if you didn't ask him
about it, and if you did, it would be interesting to this discussion to
know what the problems where.
"The Author" ?
This is open source code; there may be 27 authors of the relevant file.
In XFree86 code I wouldn't know how to find the author of a file without
looking at that file. My {limited ,mis}understanding of clean room coding
makes me wary of reading any source unless I know that its licence will
allow me to do what I wish.
This is not acceptable. You are making wild accusations, and didn't even
try to contact the relevant people. To my knowledge, Petr is the sole
author of matroxfb, and there should not have been any problem in at
least asking him about this.
OK. So I've probably been paranoid and lazy, but if the fbdev licence
had been compatible with the XFree86 one, I would have done the work.
As it is the bar was raised high enough to stop me.
Yeah, whatever, but with you asking that the fbdev drivers change their
licence, it is the same thing as the GPL zealots not liking the new
XFree86 licence.
The way to solve this is by cooperation, not by staying aloft and
pointing the finger to the opposite side.
quoted
quoted
So, for one developer at least, the reason there has been no traffic
from fbdev to XFree86 is *directly* because of the licence issue.
Yeah, but again, was it so because of a definite will on the fbdev
authors part, or because you didn't ask him ?
Isn't the aim of open source licences is to allow people to use the code
without tracking down the author and obtaining permission ?
Yes. But the aim of GPLed code is that those author give you the
permission, but also force you to give back the changes you do under the
same licence. And altough i contribute to project with the licence the
project choose, i would never choose something else as the GPL for my
own projects. If someone else wants access to the code, they can ask me
for it, and we can discuss stuff and arrive to an arrangement.
I can do that with closed source.
Well, the only reason you need to contact the author is because you want
some additional right from him, if your project was GPLed, it was no
problem.
Friendly,
Sven Luther
-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn
From: Harold L Hunt II <hidden> Date: 2004-02-02 13:13:49
Sven Luther wrote:
On Sat, Jan 31, 2004 at 01:06:23PM +0000, Andrew C Aitchison wrote:
quoted
On Sat, 31 Jan 2004, Sven Luther wrote:
quoted
On Sat, Jan 31, 2004 at 09:10:22AM +0000, Andrew C Aitchison wrote:
quoted
For several years the mga fb kernel driver has supported dual head and/or
dvi on cards which aren't supported by the XFree86 driver (unless you
use the mga_hal). I've wanted to use kernel code to add this support to XFree86, but been put off by the licence problem.
And, have you asked the mgafb driver author about this ?
You can hardly complain about lack of back traffic if you didn't ask him
about it, and if you did, it would be interesting to this discussion to
know what the problems where.
"The Author" ?
This is open source code; there may be 27 authors of the relevant file.
In XFree86 code I wouldn't know how to find the author of a file without
looking at that file. My {limited ,mis}understanding of clean room coding makes me wary of reading any source unless I know that its licence will allow me to do what I wish.
This is not acceptable. You are making wild accusations, and didn't even
try to contact the relevant people. To my knowledge, Petr is the sole
author of matroxfb, and there should not have been any problem in at
least asking him about this.
Wild accusations? How do you get wild accusations from pointing out that there "may be 27 authors of the relevant file"? If anyone is making wild accusations, it is you. Andrew simply stated the point that this is not an issue about proving whether *one* file doesn't have issues; rather, it is the issue of having to prove that *all* files do not have issues, and many of these files may be just as messy in authorship as Andrew is suggesting.
Harold
-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn
From: Dr Andrew C Aitchison <hidden> Date: 2004-02-02 14:00:05
On Sat, Jan 31, 2004 at 09:10:22AM +0000, Andrew C Aitchison wrote:
On Fri, 30 Jan 2004, Sven Luther wrote:
quoted
Yeah, that would be rather problematic, but anyway, most of the things
move from the XFree86 code to fbdev code, and most often, it is not code
that is copied, but the register information and such. It is always
easier to get specs if you are working for XFree86 than if you plan to
do some kernel driver work.
On Sat, 31 Jan 2004, Benjamin Herrenschmidt wrote:
quoted
The fact that it is mostly a one way is mostly due to the fact that the
main problem here is seeking for HW informations.
For several years the mga fb kernel driver has supported dual head and/or
dvi on cards which aren't supported by the XFree86 driver (unless you
use the mga_hal). I've wanted to use kernel code to add this support to
XFree86, but been put off by the licence problem.
On Sat, 31 Jan 2004, Sven Luther wrote:
quoted
quoted
And, have you asked the mgafb driver author about this ?
You can hardly complain about lack of back traffic if you didn't ask him
about it, and if you did, it would be interesting to this discussion to
know what the problems where.
On Sat, Jan 31, 2004 at 01:06:23PM +0000, Andrew C Aitchison wrote:
"The Author" ?
This is open source code; there may be 27 authors of the relevant file.
In XFree86 code I wouldn't know how to find the author of a file without
looking at that file. My {limited ,mis}understanding of clean room coding
makes me wary of reading any source unless I know that its licence will
allow me to do what I wish.
On Mon, 2 Feb 2004, Sven Luther wrote:
This is not acceptable. You are making wild accusations, and didn't even
try to contact the relevant people. To my knowledge, Petr is the sole
author of matroxfb, and there should not have been any problem in at
least asking him about this.
I didn't intend to make *any* accusations, and don't understand what
accusations I'm supposed to have made.
I clearly have to explain my starting position more clearly;
it is probably wrong, and almost certainly the cause of most of the
confusion, however it is how I came into this arguement, and maybe seeing
how I'm thinking will let you see that I wasn't making accusations.
My understanding of copyright/patents/plagarism (I'm vague and confused
about which this covers) is that merely by reading your document,
I am allowing the possibility that I may use that information in something
which I later write.
This is the principle behind "cleanroom" development, see
http://en.wikipedia.org/wiki/Cleanroom,
meaning 2.
If my licence to use your document doesn't allow me to do what I wish,
it is therefore better for me not to read your document.
My understanding about fbdev was that it was GPL-licenced, and that
it is *not* OK to incorporate GPL-ed code into XFree86.
Since I can't read the source code, I can't see who owns the bit I'm
interested in and can't therefore ask permission to use it under a
different licence.
I merely wished to point out that the GPL-licence *had* affected my
decision not to copy anything from fdbdev into XFree86.
Call me lazy, mis-informed, confused and paranoid, but I resent the
suggestion that I've been making accusations, wild or tame.
quoted
OK. So I've probably been paranoid and lazy, but if the fbdev licence
had been compatible with the XFree86 one, I would have done the work.
As it is the bar was raised high enough to stop me.
Yeah, whatever, but with you asking that the fbdev drivers change their
licence, it is the same thing as the GPL zealots not liking the new
XFree86 licence.
The way to solve this is by cooperation, not by staying aloft and
pointing the finger to the opposite side.
I didn't intend to ask that fbdev change their licence (although I wish
they would). I merely intended to point out that, however much the fault
was mine, the perception of the licence conflict had blocked transfer
from fbdev to XFree86.
Since Sven and Benjamin both suggested that transfer from fbdev to
XFree86 wasn't important, I thought it reasonable to relate my
experience showing that transfer in that direction was desirable and
that the GPL-licence was a hinderance.
I also realize that I have a habit of using complex and precise English.
As many people in this discussion are not native English speakers,
that is not smart, and may be why some of my intended meaning has
not got through. I apologize for this.
--
Dr. Andrew C. Aitchison Computer Officer, DPMMS, Cambridge
A.C.Aitchison@dpmms.cam.ac.uk http://www.dpmms.cam.ac.uk/~werdna
-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn
From: Sven Luther <hidden> Date: 2004-02-02 16:04:57
On Mon, Feb 02, 2004 at 08:13:45AM -0500, Harold L Hunt II wrote:
Sven Luther wrote:
quoted
On Sat, Jan 31, 2004 at 01:06:23PM +0000, Andrew C Aitchison wrote:
quoted
On Sat, 31 Jan 2004, Sven Luther wrote:
quoted
On Sat, Jan 31, 2004 at 09:10:22AM +0000, Andrew C Aitchison wrote:
quoted
For several years the mga fb kernel driver has supported dual head
and/or
dvi on cards which aren't supported by the XFree86 driver (unless you
use the mga_hal). I've wanted to use kernel code to add this support to
XFree86, but been put off by the licence problem.
And, have you asked the mgafb driver author about this ?
You can hardly complain about lack of back traffic if you didn't ask him
about it, and if you did, it would be interesting to this discussion to
know what the problems where.
"The Author" ?
This is open source code; there may be 27 authors of the relevant file.
In XFree86 code I wouldn't know how to find the author of a file without
looking at that file. My {limited ,mis}understanding of clean room coding
makes me wary of reading any source unless I know that its licence will
allow me to do what I wish.
This is not acceptable. You are making wild accusations, and didn't even
try to contact the relevant people. To my knowledge, Petr is the sole
author of matroxfb, and there should not have been any problem in at
least asking him about this.
Wild accusations? How do you get wild accusations from pointing out
that there "may be 27 authors of the relevant file"? If anyone is
making wild accusations, it is you. Andrew simply stated the point that
Ok, sorry, shouldn't have said it so, maybe it was a bit exagerated.
Still this is degenerating in GPL bashing, which will bring us nowhere.
And if you didn't make the effort to ask at least once, where will we
go. I am sure that a post to the linux-fbdev mailing list would have
solved everything, or maybe the main maintainer of the matroxfb driver ?
this is not an issue about proving whether *one* file doesn't have
issues; rather, it is the issue of having to prove that *all* files do
not have issues, and many of these files may be just as messy in
authorship as Andrew is suggesting.
Still, there is a mailing list where all linux fbdev authors are, or at
least most of them, and bitkeeper will mostly give us full history, so i
doubt it is as bad as you say.
Also, most fbdev authors have also been XFree86 contributors in the
past, so i don't really know what the problem is here.
Friendly,
Sven Luther
-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn
From: Sven Luther <hidden> Date: 2004-02-02 16:11:44
On Mon, Feb 02, 2004 at 01:59:54PM +0000, Dr Andrew C Aitchison wrote:
On Sat, Jan 31, 2004 at 09:10:22AM +0000, Andrew C Aitchison wrote:
quoted
On Fri, 30 Jan 2004, Sven Luther wrote:
quoted
Yeah, that would be rather problematic, but anyway, most of the things
move from the XFree86 code to fbdev code, and most often, it is not code
that is copied, but the register information and such. It is always
easier to get specs if you are working for XFree86 than if you plan to
do some kernel driver work.
On Sat, 31 Jan 2004, Benjamin Herrenschmidt wrote:
quoted
The fact that it is mostly a one way is mostly due to the fact that the
main problem here is seeking for HW informations.
For several years the mga fb kernel driver has supported dual head and/or
dvi on cards which aren't supported by the XFree86 driver (unless you
use the mga_hal). I've wanted to use kernel code to add this support to
XFree86, but been put off by the licence problem.
On Sat, 31 Jan 2004, Sven Luther wrote:
quoted
quoted
quoted
And, have you asked the mgafb driver author about this ?
You can hardly complain about lack of back traffic if you didn't ask him
about it, and if you did, it would be interesting to this discussion to
know what the problems where.
On Sat, Jan 31, 2004 at 01:06:23PM +0000, Andrew C Aitchison wrote:
quoted
"The Author" ?
This is open source code; there may be 27 authors of the relevant file.
In XFree86 code I wouldn't know how to find the author of a file without
looking at that file. My {limited ,mis}understanding of clean room coding
makes me wary of reading any source unless I know that its licence will
allow me to do what I wish.
On Mon, 2 Feb 2004, Sven Luther wrote:
quoted
This is not acceptable. You are making wild accusations, and didn't even
try to contact the relevant people. To my knowledge, Petr is the sole
author of matroxfb, and there should not have been any problem in at
least asking him about this.
I didn't intend to make *any* accusations, and don't understand what
accusations I'm supposed to have made.
I clearly have to explain my starting position more clearly;
it is probably wrong, and almost certainly the cause of most of the
confusion, however it is how I came into this arguement, and maybe seeing
how I'm thinking will let you see that I wasn't making accusations.
My understanding of copyright/patents/plagarism (I'm vague and confused
about which this covers) is that merely by reading your document,
I am allowing the possibility that I may use that information in something
which I later write.
This is the principle behind "cleanroom" development, see
http://en.wikipedia.org/wiki/Cleanroom,
meaning 2.
If my licence to use your document doesn't allow me to do what I wish,
it is therefore better for me not to read your document.
My understanding about fbdev was that it was GPL-licenced, and that
it is *not* OK to incorporate GPL-ed code into XFree86.
Since I can't read the source code, I can't see who owns the bit I'm
interested in and can't therefore ask permission to use it under a
different licence.
I merely wished to point out that the GPL-licence *had* affected my
decision not to copy anything from fdbdev into XFree86.
Call me lazy, mis-informed, confused and paranoid, but I resent the
suggestion that I've been making accusations, wild or tame.
Well, i seriously doubt that reading the first lines of a file would
contaminate you, after all you could use head to look at them or
something. Also, you could have written to linux-fbdev mailing list if
you were interested, or even have asked on devel@xfree86.org, and those
with interest in fbdev matters would have responded to you.
quoted
quoted
OK. So I've probably been paranoid and lazy, but if the fbdev licence
had been compatible with the XFree86 one, I would have done the work.
As it is the bar was raised high enough to stop me.
Yeah, whatever, but with you asking that the fbdev drivers change their
licence, it is the same thing as the GPL zealots not liking the new
XFree86 licence.
The way to solve this is by cooperation, not by staying aloft and
pointing the finger to the opposite side.
I didn't intend to ask that fbdev change their licence (although I wish
they would). I merely intended to point out that, however much the fault
was mine, the perception of the licence conflict had blocked transfer
from fbdev to XFree86.
Well, i believe this was not the only issue involved. And i guess that
if the XFree86 Project was asking the fbdev people that they move to a
dual licence or something such, in order to better share information and
code between the fbdev drivers and the XFree86 drivers, then
undoubtfully this would have happened, with the natural problem of
searching for non-active authors of past code. I believe that Benjaming
Herrenschmidt raising its concern about driver code consecutive to the
announcement of the XFree86 licence could be the natural place to start
such an agreement.
Since Sven and Benjamin both suggested that transfer from fbdev to
XFree86 wasn't important, I thought it reasonable to relate my
experience showing that transfer in that direction was desirable and
that the GPL-licence was a hinderance.
Ok.
I also realize that I have a habit of using complex and precise English.
As many people in this discussion are not native English speakers,
that is not smart, and may be why some of my intended meaning has
not got through. I apologize for this.