gnuplot-beta — Developers and beta testers of new gnuplot code

Re: gnuplot 4.1 source

[Hans-Bernhard B. <br...@ph...>]

Darrell Kindred wrote:
> $ cvs -z3 -d:pserver:ano...@cv...:/cvsroot/gnuplot co -P gnuplot
> can't create temporary directory /tmp/cvs-serv6605
> Too many links


> This appears to be a server-side error, 

What do you deduce that from?  Did you try cvs -t -z3 ... to get it to 
report the entire conversation your client has with the server?

 > and it seems to be limited to the gnuplot repository.

Well, somebody should report that to SF.net then ('Help' in the upper 
bar, 'Get Support' in the lower, then follow the instructions to posting 
a support request).  I suggest you do it.  We, as the project admins, 
have no more direct control over those CVS servers than anonymous users 
do --- this has to be handled by SF.net staff.

What's left for a 4.2 release? [feature requests]

[Ethan A M. <merritt@u.washington.edu>]

Almost 3 years ago, we had a thread titled 
"What's left for a Version 4.0 release?"
It was useful in guiding the long process of readying the 4.0 official
release.  I am hoping that a similar thread will kick off the 4.2
release process.  I'm also hoping that this time it will take less
than a year to complete the process. We'll see.

Here is my evaluation of the state of the code base, and the
collection of bugs/patches/feature-requests on SourceForge.
I've split it into several parts:
   core code
   drivers
   patches
   feature requests

Feature Requests
================

Summary:
A suggestion with no actual code does not count for as much
as a real patchset.  That said, there are a few good ideas
on the SourceForge site that might be worth looking at. If the
implementation is easy enough, we might consider working on them
for 4.2. The following look like plausible candidates to me:

#1376595         Class and name strings of X Window
#1117724         [fit] access to resulting chisquare
#1078852         Flexible usage of geographic coordinates (ex. DD:MM:SS)
#674055          mouse coordinate read-out in multiplot mode



-- 
Ethan A Merritt
Biomolecular Structure Center
University of Washington, Seattle 98195-7742

What's left for a 4.2 release? [patches]

[Ethan A M. <merritt@u.washington.edu>]

Almost 3 years ago, we had a thread titled 
"What's left for a Version 4.0 release?"
It was useful in guiding the long process of readying the 4.0 official
release.  I am hoping that a similar thread will kick off the 4.2
release process.  I'm also hoping that this time it will take less
than a year to complete the process. We'll see.

Here is my evaluation of the state of the code base, and the
collection of bugs/patches/feature-requests on SourceForge.
I've split it into several parts:
   core code
   drivers
   patches
   feature requests

Patchsets
=========

Summary:
Although the current collection of patchsets on SourceForge
represents a lot of work by a lot of people, I don't think there
is much there that cries out for inclusion in a 4.2 release.
Here's my attempt at sorting them:

I'm willing to spend time to get it into 4.2
==================================================
1267434         wxWidgets terminal
1218873         Change text rotating angle from int to float
908040          Break PostScript prologue file out of post.trm


Maybe, if someone else does the work :-)
==================================================
1244775         Isolines optional on datafile splots
1199186         strftime and strptime
1077726         true depth ordering for pm3d plots


Good idea, but not for 4.2
==================================================
1252232         Allow UTF-8 input for normal postscript Type1 fonts
1252215         Font kerning in postscript terminal
1143563         Place arbitrary objects in plots
1196485         OpenGL terminal driver
1105717         Add '=' special file to reread '-' inline data


I cannot judge these
=================================================
1398474         a driver for GD.pm
1384525         Small fixes to gd terminals
1294507         Fitting using CERN Minuit routines
1248308         Rexx scripting support
1233137         Font control in tkcanvas
1289725         handling double quotes in gnuplot-mode


Reject for now; revisit later
=============================
My evaluation is that the rest of these should be rejected,
and the core ideas revisited later if there is still a need
after the code destined for 4.2 has settled down.
Note that several of these involve the clipping code,
which I have separately listed as a release-critical issue.
I just don't think that these particular patches are the way
to get there, even though some of them were my own suggestions.

1364114         Extension of Ethans patch: Clipping in Postscript terminal
1356114         set term ... size <foo>,<baz>
1353539         Fix clipping for plots with canvas size != 1,1
1329098         New clipping code
1328103         Fix BoundingBox for large postscript plots
1321476         Draw zeroaxis at specified positions
1318546         Add 'offset' and 'backhead' to 'set arrow'
1262281         Povray terminal driver
1251204         Real 3d perspectivic look to splot
1206823         Boxed text labels
1193448         Use allterm.h for texinfo documentation
1185346         Autoscaling with constraint on final range.
1123355         Support a header file in epslatex terminal
1104264         Fix buggy clipping of arrows in large splots
1052938         Fix incorrect alignment of rotated multiline text
1044573         change density of generated tics
1030055         add user-specified colors to emf.trm
1027032         Connect gnuplot_x11 to exterior application window
1024305         pie graphs (work in progress)
982765          Patch for AI (Adobe Illustrator) term
936695          Some new logos
760421          plot smooth, with set log y
633724          new terminals:Python/Tk and Ruby/Tk
632289          new dgrid3d options: scaling and range
588805          external functions via plugins
551439          plot ... index by name instead of number





-- 
Ethan A Merritt
Biomolecular Structure Center
University of Washington, Seattle 98195-7742

What's left for a 4.2 release? [drivers]

[Ethan A M. <merritt@u.washington.edu>]

Almost 3 years ago, we had a thread titled=20
"What's left for a Version 4.0 release?"
It was useful in guiding the long process of readying the 4.0 official
release.  I am hoping that a similar thread will kick off the 4.2
release process.  I'm also hoping that this time it will take less
than a year to complete the process. We'll see.

Here is my evaluation of the state of the code base, and the
collection of bugs/patches/feature-requests on SourceForge.
I've split it into several parts:
   core code
   drivers
   patches
   feature requests


Drivers
=3D=3D=3D=3D=3D=3D=3D

Summary:
The major changes since 4.0 involved re-working the PostScript
driver and its derivatives.  I think we made a mistake in leaving
old drivers in the 4.0 release, and should aim for simply removing
obsolete drivers in 4.2.  Consider how many queries on the newsgroup
turn out to be from people whose gnuplot 4.0 was built with the
1995-era png driver rather than with libgd.

Issues:
 - New drivers
   + aquaterm is already in cvs.  I have some reservations about
     making it a default, however, because it does not support all
     the features of the x11 terminal.  IMHO either the remaining
     features should be added, or x11 should remain the default
     terminal even on systems that support aqua.
   + Timoth=C3=A9e's wxWidgets driver is looking very solid, and should
     go into cvs
   + I'm glad to see there is ongoing work on an OpenGL terminal,
     but I don't see much chance anything will be ready for 4.2

 - Old but not obsolete drivers
   + Windows is the 800 pound gorilla here.  It has fallen behind=20
     the other interactive drivers in capability, and it does not
     seem to have a champion on the development team to bring it up
     to par.  Can we move to the wxWidgets driver instead, and=20
     retain the existing windows driver only for older systems that
     don't support wxWidgets?

 - Obsolete drivers (any machine that can run these has a better
   alternative driver). Should we remove them from the tree altogether?
   + openstep/next
	I'm not an expert on the history of Apple O/S releases, but
	it is my understanding that any machine these were relevant
	to could now run the normal postscript driver for printer
	output and x11 for interactive output.  I used to have
	a Next machine, and it did work for both postscript and
	x11. But it's long gone.
   + ggi
	Does not link properly against the ggi+xmi libraries currently
	available on the web.
	The basic ggi.trm works for me, but does not do anything useful
	that cannot be better done with another terminal type.
	As best as I can make out, gnuplot's xmi code was written to match
	the SCO implementation, which is for all practical purposes
	dead. And given SCO's claims of proprietary code, it is probably
	not usable anyhow. =20
   + iris4d
	I work in a lab that has used SGI workstations for at least the
	last 12 years.  *No one* uses the iris4d driver, and they never
	have.  You would think that the attraction of a separate driver
	for 3D workstations would be that it supports 3D... but it doesn't.
   + ai
	So far as I know, "set term post level1" is sufficient to=20
	generate output compatible with Adobe Illustrator (ai).
	If not, I'd rather fix post.trm than try to support the ancient
	ai.trm.

Other dubious drivers:
   + svgalib
	I have never managed to get this one to work right, and it's
	a blatant security hole. But I know that some people use it,
	because we get occasional bug reports [which don't get fixed].
	Anyone know of a way to provide a linux console terminal=20
	driver that doesn't have these problems?

Bugs:
Most of the bug reports on SourceForge are driver-specific.
Some are ancient, and some can probably be ignored.

windows:
#1413021        [Wgnuplot] pause 1;reread; blocks interaction
#1232950        -persist option does not work in Version 4.0.0 (Windows)
#982293         can't print color in win32 gnuplot 4.0
#561418         (MS Windows) 100% CPU Usage during pause
#233405         WGNUPL32.EXE crashes when printing directly (Win 95)

x11:
#1376604        misleading "noraise" resource; resource list in doc
#1305877        x11: cursor position in clipboard N/A in Qt/KDE applications
#997481         [autoconf] name transformation of gnuplot_x11

TeX:
#1377786        kpsexpand/kpsewhich from tetex are called unconditionally
#1367060        epslatex generates ambiguous TeX code
#1356987        erroneous eepic output for ylabel
#1355374        term tkcanvas perltk broken

other:
#1224391        Linux console corrupt screen after switch
#1039296        PDFlib Lite 6 output is broken for pipes, stdout
#596701         tkcanvas perltk problem

=2D-=20
Ethan A Merritt
Biomolecular Structure Center
University of Washington, Seattle 98195-7742

What's left for a 4.2 release? [core code]

[Ethan A M. <merritt@u.washington.edu>]

Almost 3 years ago, we had a thread titled 
"What's left for a Version 4.0 release?"
It was useful in guiding the long process of readying the 4.0 official
release.  I am hoping that a similar thread will kick off the 4.2
release process.  I'm also hoping that this time it will take less
than a year to complete the process. We'll see.

Here is my evaluation of the state of the code base, and the
collection of bugs/patches/feature-requests on SourceForge.
I've split it into several parts:
   core code
   drivers
   patches
   feature requests

Core code
===========

Summary:
Either the cvs code is very stable, or no one is giving it a workout.
I think it's the former.  Most of the recent bug reports against the
cvs version have turned out to be old problems that date back to 
3.7 or earlier.  The only major sore point that I see is the
clipping code. More on that later.

Issues:

 - How many of the new options should be enabled by default?
   Should any of them continue to be marked EXPERIMENTAL?

 - Can we confirm that the code builds and runs on amiga?  VMS?
   If it doesn't even build, do we try to fix it or do we just stop
   claiming to support those platforms?

 - The binary file code is ugly, ugly, ugly.  But it works.
   If we hold up a release until all the code is beautiful, it will
   never happen.

 - Feature requests, patches
   I've started a separate thread for these.

Known bugs:

 - Clipping, particularly of arrows and filled polygons, is still
   not done cleanly and therefore some terminals crash on reasonable
   input.  
   This is the only known bug that I consider release-critical.

 - Treatment of "missing" data as opposed to "invalid" data.
   I think this is more a matter of poor documentation than an actual
   bug, but I could be wrong.  In any event, it's not fixable until
   someone can point to an unambiguous statement of what *should*
   happen in both cases, and a reproducible example that doesn't act
   that way. SF bugs #775810 #918793 #969322 #1403945

 - Miscellaneous open bug reports on SourceForge are related to
   the core code rather than to individual terminals.
   The list is actually quite short:

   #1408955        ytics have wrong date with ydata time and "using 2"
   #1363641        dgrid3d / xrange interaction
   #1346814        cannot unset mxtics for time data
   #1184989        set key below noautotitles; plot x -- crash
   #1182499        set terminal doesn't automatically close output
   #1158281        plotting a ternary function with undefined values
   #1107709        plot [-1:1] x is plotted with asymmetric y-axis
   #1058117        Legend Disappears on Some Plots
   #1042785        reversed axis range breaks filled curves
   #1039309        Strange xtics on 'time' plot
   #1024394        bug on time/data format
   #1004754        Tics and grid slightly outside border.
    #992528        `set offset` can break `with filledcurves`
    #634506        plotting a subrange of a contour plot


-- 
Ethan A Merritt
Biomolecular Structure Center
University of Washington, Seattle 98195-7742

Re: gnuplot 4.1 source

[Darrell K. <dk....@qu...>]

Ethan Merritt wrote:
> Yes.  Download it from cvs.
> The instructions on SourceForge are complete and correct
> as to how you can do this.

I too am having no luck checking out gnuplot code via CVS
today, though the error is different than Winnie reported:

$ cvs -d:pserver:ano...@cv...:/cvsroot/gnuplot login
Logging in to :pserver:ano...@cv...:2401/cvsroot/gnuplot
CVS password: 
$ cvs -z3 -d:pserver:ano...@cv...:/cvsroot/gnuplot co -P gnuplot
can't create temporary directory /tmp/cvs-serv6605
Too many links
$ 

This appears to be a server-side error, and it seems to be
limited to the gnuplot repository.  I'm able to check out
other sourceforge projects successfully, such as

   cvs -d:pserver:ano...@cv...:/cvsroot/itools-htmlarea login
   cvs -z3 -d:pserver:ano...@cv...:/cvsroot/itools-htmlarea co -P htmlarea

Googling indicates that others have seen similar errors from
anonymous cvs servers at sourceforge and elsewhere in the past.

If any of the maintainers are able to get this resolved,
and/or if anyone can send me a relatively recent 4.1
snapshot, I'd greatly appreciate it.

Thanks,

Darrell
(habitual gnuplot abuser)

Re: [Fwd: [ILUG] gnuplot datafile with rows]

[Ethan A M. <merritt@u.washington.edu>]

On Thursday 26 January 2006 09:26 am, Ethan Merritt wrote:
> 
> gnuplot 4.1  (Note version number!)
> 
> # Power along x; Day on y
> # This *almost* works, but I'm having trouble getting it
> # to skip the non-numeric data in the 1st column
> # I think there's a bug in the program :-(

I have fixed this bug and put the fix into cvs.
The corrected version should be user-accessible by the
weekend (SourceForge propagation times are a bit unpredictable).

> #
> set style data lines
> plot 'datafile' matrix using 1:3 every 1:999:1:0, \
>         '' matrix using 1:3 every 1:999:1:1, \
>         '' matrix using 1:3 every 1:999:1:2, \
> 
> The "matrix using 1:3"  causes it to plot rows
> The "every 1:999:<start column>:<start row>"
> selects which row and which column to start with.
> Row and column numbering starts at 0, so putting a
> 1 for <start column> should skip the first column of data.

This set of commands should now work on your input file.


-- 
Ethan A Merritt
Biomolecular Structure Center
University of Washington, Seattle 98195-7742

Re: multiline xtics

[Johannes Z. <joh...@ze...>]

On Thu, Jan 26, 2006 at 03:43:05PM +0100, Hans-Bernhard Broeker wrote:
> Johannes Zellner wrote:
> >how do I get multiline xtics?
> >
> >set xtics ("bla\nfred" 1)
> >
> >doesn't work.
> 
> Then you must be using a different gnuplot than I am (4.0.0 release 
> binary on Win32 for this test) --- because that quite definitely *does* 
> work.  You have to remember to use "" quotes, not '' ones, though.

well, actually it does work, you're right! But the second line doesn't
appear (in X11) as no space is reserved. Try

    set xtics ("fred\nhello" 1)
    plot x*x

--> "hello" does not appear.

    set xlabel "hi"
    rep

--> "hello" appears, as the xlabel reserves more space.

-- 
Johannes

Re: [Fwd: [ILUG] gnuplot datafile with rows]

[Ethan M. <merritt@u.washington.edu>]

On Thursday 26 January 2006 01:36 am, Lars Hecking wrote:
>  Can anyone help with this?
>
> ----- Forwarded message from Karine Dionne <k.d...@ja...>
>
> I am new to Gnuplot and to Linux as well and I am having a bit a
> struggle right now.  Here is the format of my datafile:
>
> Day	Power1	Power2	Power3	Power4	...
> Day1	y1	y1	y1	y1	...
> Day2	y2	y2	y2	y2	...
> Day3 	y3	y3	y3	y3	...

gnuplot 4.1  (Note version number!)

#Day along x;  Power reading on y
# (Note: You must uncomment the first line of your data file
#             if you want it to be used for generating key titles)
#
set style data lines
set key autotitle columnhead
plot 'datafile'  using 2:xtic(1), '' using 3, '' using 4, '' using 5


# Power along x; Day on y
# This *almost* works, but I'm having trouble getting it
# to skip the non-numeric data in the 1st column
# I think there's a bug in the program :-(
#
set style data lines
plot 'datafile' matrix using 1:3 every 1:999:1:0, \
        '' matrix using 1:3 every 1:999:1:1, \
        '' matrix using 1:3 every 1:999:1:2, \

The "matrix using 1:3"  causes it to plot rows
The "every 1:999:<start column>:<start row>"
selects which row and which column to start with.
Row and column numbering starts at 0, so putting a
1 for <start column> should skip the first column of data.
Unfortunately this is where the bug creeps in; it will skip
the first column if the column contains a number, but 
complains "Bad number in matrix" if the contents are
non-numeric.   I'll see if I can fix that bug this weekend,
but for now you will have to pre-filter the input to remove
the first column using some appropriate utility like
sed, awk, perl, or egrep.   Here I use sed to strip out
leading non-blank characters from each line:

plot "< sed -e 's/^[^ ]*//' datafile" using 1:3 every 1:999:0:0, \
 "" matrix using 1:3 every 1:999:0:1, \
 "" matrix using 1:3 every 1:999:0:2, \
        etc for all the rows

If you want to hit all of the rows together in one command,
leave out the "every" clause:
  plot "< sed -e 's/^[^ ]*//' datafile" using 1:3


So yeah, it's possible.
But your specific request seems to have uncovered a bug
in the "every" behaviour, so you will need to use some
annoying work-around like my sed command above until
that's fixed.

		Ethan



>
> "DayX" is actually a date string, but that is not my problem
> for now.  Each column is a power measurement  given by a specific
> powermeter.  There are two types of plots I need to do.
>
> To begin, I want to compare the performance of each powermeter on a
> single plot.  So, I would need to generate an X vector for the
> powermeters' IDs (from 1 to 32 in steps of 1) and I don't know how to
> do that.  I would also need to be able to tell gnuplots to proccess
> rows, not columns. I browsed the internet and the mailing lists, I
> found one thread about it. It was in March 2000 and it says that you
> have to reformat your file in another file and then plot the columns.
>
> There are 2 reasons why I don't want to do that.  First of all, this
> is a log file that will become very big with time.  I am not
> interested in having to keep two copies of them.  I could keep only
> the transposed copy, but this solution is not possible because of the
> second type of plot I need.  I want to track the changes with time of
> the power meters.  I will use the date string as the X axis and the
> other columns as Y.  So, one way or another, I have to access rows of
> data for one of my two plots.
>
> I would like to know if there is a possibility to do that with
> gnuplot or if I should consider using something else that could do
> the job.  I with to find something efficient and as straight forward
> as possible.  I hope I am not asking for to much.
>
> Thanks for your time and expertise.
>
> Karine
>
> ----- End forwarded message -----
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through
> log files for problems?  Stop!  Download the new AJAX search engine
> that makes searching your log files as easy as surfing the  web. 
> DOWNLOAD SPLUNK!
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121
>642 _______________________________________________
> gnuplot-beta mailing list
> gnu...@li...
> https://lists.sourceforge.net/lists/listinfo/gnuplot-beta

-- 
Ethan A Merritt
Biomolecular Structure Center
University of Washington, Seattle WA

Re: multiline xtics

[Hans-Bernhard B. <br...@ph...>]

Johannes Zellner wrote:
> how do I get multiline xtics?
> 
> set xtics ("bla\nfred" 1)
> 
> doesn't work.

Then you must be using a different gnuplot than I am (4.0.0 release 
binary on Win32 for this test) --- because that quite definitely *does* 
work.  You have to remember to use "" quotes, not '' ones, though.

multiline xtics

[Johannes Z. <joh...@ze...>]

Hello,

how do I get multiline xtics?

set xtics ("bla\nfred" 1)

doesn't work.


-- 
Johannes

Good publicity !

[<tim...@en...>]

Hi gnuplot developpers !

The LWN.net issue of the last week (made freely available today)
contains a good article about software for scientific publications :
"Using open-source tools for documenting research". The author, Carl
Bolduc, makes a quick inventory of scientific software which can replace
the expensive ones commonly used by lots of researchers. The article
focuses on publications, not on computations or data processing. You
will happily note that gnuplot is the software chosen for producing high
quality plots.

Quote :
"Gnuplot
Continuing in your path to build an open-source research writing box,
you need a powerful tool to generate plots from your precious experiment
results. That is where Gnuplot enters the scene, with its almost
limitless possibilities. Gnuplot is a command-line plotting utility with
easy to learn commands that enable you to create high quality 2D and 3D
plots suitable for scientific publications. It can output LaTeX and EPS
code which can be inserted in your LaTeX documents. You can check out
this demo that shows the wide variety of Gnuplot's capabilities."

The url : http://lwn.net/Articles/167317/

LWN.net is a weekly worthwile information source, specialized in free
software and Linux.

Regards,

Timoth=E9e Lecomte

[Fwd: [ILUG] gnuplot datafile with rows]

[Lars H. <lhe...@us...>]

 Can anyone help with this?

----- Forwarded message from Karine Dionne <k.d...@ja...> -----

Date: Wed, 25 Jan 2006 09:50:07 -1000 (HST)
From: Karine Dionne <k.d...@ja...>
To: il...@li...
Subject: [ILUG] gnuplot datafile with rows

Hi,

I am new to Gnuplot and to Linux as well and I am having a bit a struggle 
right now.  Here is the format of my datafile:

#Day	Power1	Power2	Power3	Power4	...
Day1	y1	y1	y1	y1	...
Day2	y2	y2	y2	y2	...
Day3 	y3	y3	y3	y3	...

"DayX" is actually a date string, but that is not my problem 
for now.  Each column is a power measurement  given by a specific 
powermeter.  There are two types of plots I need to do.

To begin, I want to compare the performance of each powermeter on a single 
plot.  So, I would need to generate an X vector for the powermeters' IDs 
(from 1 to 32 in steps of 1) and I don't know how to do that.  I would 
also need to be able to tell gnuplots to proccess rows, not columns. I 
browsed the internet and the mailing lists, I found one thread about it. 
It was in March 2000 and it says that you have to reformat your file in 
another file and then plot the columns.

There are 2 reasons why I don't want to do that.  First of all, this is a 
log file that will become very big with time.  I am not interested in 
having to keep two copies of them.  I could keep only the transposed copy, 
but this solution is not possible because of the second type of plot I 
need.  I want to track the changes with time of the power meters.  I will 
use the date string as the X axis and the other columns as Y.  So, one way 
or another, I have to access rows of data for one of my two plots.

I would like to know if there is a possibility to do that with gnuplot or 
if I should consider using something else that could do the job.  I with 
to find something efficient and as straight forward as possible.  I hope I 
am not asking for to much.

Thanks for your time and expertise.

Karine

----- End forwarded message -----

Re: Thoughts about gnuplot license and copyrights

[Hans-Bernhard B. <br...@ph...>]

Ethan Merritt wrote:

> Only the license needs to persist.  A permanent license remains valid
> even if the original licensing body no longer exists.

The problem is that AFAIK copyright falls under civil law, i.e. if 
somebody decided to violate the license after the copyright holder(s) 
have become terminally unavailable, there's nobody left who could sue 
them. A license without a legal body holding it would become 
unenforcable, and that cannot ever be a good thing.

The more immediate problem with the gnuplot license as-is, is that after
the copyright holders are gone, nobody can ever make an official release
again. That's a "single point of failure" problem, and really needs to
be addressed. The GPL addresses it by granting all licensees full
distribution privileges, other license (you've mentioned PINE), by
relying on an effectively immortal copyright holder.

Re: gnuplot 4.1 source

[Dmitri A. S. <das...@gm...>]

Ethan Merritt wrote:
> On Friday 20 January 2006 05:47 am, Winnie wrote:
> 
>>hi,
>>I'd love to use gnuplot 4.1 but I can't download the source code via
>>cvs, and my university computer help disk cannot resolve the problem.
>>I wonder is there any way I can obtain a copy of the version 4.1
>>source code?
> 
> 
> Yes.  Download it from cvs.
> The instructions on SourceForge are complete and correct
> as to how you can do this.
> 

cvs has been flaky for the past few days (I cannot get a cvs dump
either):

[dima@tumbleweed cvs-dump]$  cvs -d:pserver:ano...@cv...:/cvsroot/gnuplot login
Logging in to :pserver:ano...@cv...:2401/cvsroot/gnuplot
CVS password:
cvs [login aborted]: reading from server: Connection reset by peer

(the error messages were different)

I tried from 3 different location (different ISPs) w/ the same effect.

That might be the reason for "I can't download the source code via
cvs".

Sincerely,

Dmitri.
-- 

Re: Thoughts about gnuplot license and copyrights

[Ethan M. <merritt@u.washington.edu>]

On Monday 23 January 2006 07:10 am, Hans-Bernhard Broeker wrote:
> Ethan A Merritt wrote:
> > I conclude that you would need on the order of 25 signatures to
> > fully satisfy the FSF, if that is your goal. But that specific
> > criterion would only be triggered if you were attempting to
> > transfer copyright to the FSF, which is rather extreme.
>
> You would need equally many signatures regardless of what kind of
> legal body you assign the copyright to.

Maybe; I'm not sure. But I think the issue of copyright is sidetracking
the main point under discussion.

It is not necessary to transfer copyright in order to change a license.
End users have access to the code, with certain restrictions, because
of the license.  The current holder of the copyright is not all that
relevant if the license grants permanent rights to others.  If ther
code were relicensed following the model of  GPL or BSD or Pine,
development could continue even if the original copyright holders
were to drop out of sight.

What is true, however, is that in order to change the license we
would have to get the explicit approval of Thomas, and it would
certainly be safest to get explicit approval from as many other
names on the list as possible.

I point out again that individual contributors can make their
contributions available under multiple licenses.  Timoth=E9e, for
example, can mark his wxWidgets code as being licensed under
the GPL if he wishes. In this case he must *also*  license it under
gnuplot's current license (or possibly LGPL) in order for gnuplot
to use it.  But dual-licensing would allow 3rd parties to re-use
the driver code in other projects regardless of the course of
gnuplot development.

> * Get ready to face the music.  Sign over the copyrights to some
> legal body.  Either to an existing institution, or an individual or
> to a foundation to be created for the purpose.

You have left out the most expedient option, IMHO:

3)  Leave the copyright as is, but change the license.

=2D-=20
Ethan A Merritt
Biomolecular Structure Center
University of Washington, Seattle WA

Re: gnuplot 4.1 source

[Ethan M. <merritt@u.washington.edu>]

On Friday 20 January 2006 05:47 am, Winnie wrote:
> hi,
> I'd love to use gnuplot 4.1 but I can't download the source code via
> cvs, and my university computer help disk cannot resolve the problem.
> I wonder is there any way I can obtain a copy of the version 4.1
> source code?

Yes.  Download it from cvs.
The instructions on SourceForge are complete and correct
as to how you can do this.

-- 
Ethan A Merritt
Biomolecular Structure Center
University of Washington, Seattle WA

Re: Thoughts about gnuplot license and copyrights

[Hans-Bernhard B. <br...@ph...>]

Ethan A Merritt wrote:

> I conclude that you would need on the order of 25 signatures to fully
> satisfy the FSF, if that is your goal. But that specific criterion
> would only be triggered if you were attempting to transfer copyright
> to the FSF, which is rather extreme.

You would need equally many signatures regardless of what kind of legal 
body you assign the copyright to.   Whether that's the FSF, a 
to-be-founded gnuplot foundation, a single person, or something 
completely different; it doesn't change a thing.  It's mostly 
straightforward application of copyright law that dictates this. 
Copyright have to be assigned explicitly, in writing, to be upheld in
court.

Some people here seem to assume that the FSF is making their version of 
such a process needlessly complicated, and that it must surely be 
possible to do it a lot easier than that.  I doubt that.  I don't 
believe any of us can out-lawyer Prof. Eben Moglen.  Not even RMS, for 
that matter.

Which leaves us with basically two options:

* Don't wake a sleeping lion.  I.e. keep on going like we've been doing 
ever since about the days of the pre-3.6 beta source tarballs, under the 
optimistic assumption that nobody will feel an urge to actually sue us.

* Get ready to face the music.  Sign over the copyrights to some
legal body.  Either to an existing institution, or an individual or to a 
foundation to be created for the purpose.

Possible benefits of the latter approach would be that such an entity 
could properly join SF.net's project donation system to raise some 
funds, towards paying for stuff like DNS fees, registering a trademark 
and maybe getting back stolen domain names.

Re: Thoughts about gnuplot license and copyrights

[<tim...@en...>]

Ethan A Merritt wrote:

>On Saturday 21 January 2006 05:41 pm, Timoth=E9e Lecomte wrote:
> =20
>
>>- the FSF has a strong policy regarding copyrights : it asks for signed
>>paper for each important contributor
>>   =20
>>
>The current front page for gnuplot document lists 23 contributors
>in addition to the two original copyright holders.
> =20
>
>>What do you conclude ?
>>   =20
>>
>I conclude that you would need on the order of 25 signatures to
>fully satisfy the FSF, if that is your goal.
>But that specific criterion would only be triggered if you were
>attempting to transfer copyright to the FSF, which is rather extreme.
> =20
>
It's not my goal, and I completely agree with the "extreme" adjective !
I just wanted to confirm what Daniel J Sebald explained about the
necessity to have some legal proof if we want to do something to the
current license, such as a written letter from Thomas Williams.

Re: Thoughts about gnuplot license and copyrights

[<tim...@en...>]

Timoth=E9e Lecomte wrote:

>Daniel J Sebald wrote:
>
> =20
>
>>Timoth=E9e Lecomte wrote:
>>
>>   =20
>>
>>>I understand your point about defining legally this group. As Ethan
>>>explained, it's what the pine license does, and in this case the group
>>>which authorizes releases is the University of Washington. You can not=
e
>>>that there are major differences between pine and gnuplot : 'pine' is =
a
>>>trademark (as 'firefox' is for an other example), and there is a
>>>powerful and long-standing legal entity behind pine, the university.
>>>
>>>To remain pragmatic, I think that we can't reach the same conditions f=
or
>>>gnuplot.
>>>     =20
>>>
>>Why not?  I know the common belief now days is that only "big"
>>entities have or can afford rights, representation, etc.  But the law
>>isn't restricted to them.  In fact, a lot of these legal concepts
>>probably were originally intended to protect the individual.
>>   =20
>>
>
>Ok, it is possible, but personnaly I don't want
> =20
>
Wow, I realize that I have forgotten to finish my sentence. I meant to
say that legal protections remain complicated to handle. But I agree
with you, it's feasible.

Re: Thoughts about gnuplot license and copyrights

[<tim...@en...>]

Daniel J Sebald wrote:

> Useful list below, thanks.  Short comments within:
> Timoth=E9e Lecomte wrote:
>
>> I have wondered how things go for other projects. Indeed, instead of
>> trying to resolve the gnuplot's license disadvantage ourselves, we can
>> get the inspiration from others (that's not copyrighted !). So I have
>> made a compilation of information on famous scientific open-source
>> software. Let's see :
>>
>>
>> * octave ( numerical computations): GPL, "octave" doesn't seem to be a
>> trademark, octave is a GNU product, so it is proctected by the Free
>> Software Foundation, which takes good care of copyrights (see
>> http://www.gnu.org/prep/maintain/maintain.html  :  the FSF asks for a
>> plain paper from each important contributor that gives his copyright t=
o
>> the FSF, as for each GNU software)
>
>
> We could probably talk to John Eaton about more specifics if we
> wanted.  By plain paper you mean something with signatures?  (And
> notarized, I'd argue.)

With signatures, yes. Here is the relevant part of the link above :

    "If you maintain an FSF-copyrighted package certain legal procedures
are required when incorporating legally significant changes written by
other people.     This ensures that the FSF has the legal right to
distribute the package, and the standing to defend its GPL-covered
status in court if necessary.

    Before incorporating significant changes, make sure that the person
who wrote the changes has signed copyright papers and that the Free
Software  Foundation has received and signed them. We may also need a
disclaimer from the person's employer."

>> * Maxima (computer algebra system) : GPL, no trademark ("maxima" is
>> already the trademark of a fishing lines company),
>
>
> Doesn't have any bearing on a software license as fishing equipment is
> unrelated to software.


>> * Axiom (computer algebra system ) : BSD-style, no trademark ("axiom" =
is
>> already the trademark of an automotive company),
>
>
> Again, there can be multiple trademarks if the two companies are in
> unrelated fields.
>
> [snip]
>
>> * amarok ( music player ): GPL, no trademark, copyright is protected b=
y
>> the KDEev as it is part of KDE
>
>
> Also the name of a Mike Oldfield album.  :-)


Ok, I didn't know that there can be multiple trademarks for different
activities.

>
>> no organization behind.
>> Maxima comes from DOE-Macsyma, written by the department of Energy (as
>> far as I understand) and was officially donated to the community under
>> the GPL.
>
> Not according to the letter.  It apparently isn't GPL but says if the
> derivative work is GPL then Maxima's license must also accompany the
> GPL license.

Indeed the letter doesn't say it has to be released under the GPL. I
would say that it enumerates a list of conditions and says that Willam
Shelter has the right to sublicense the modified code he wants to
distribute. Derivated work stands here for the work done from the
original copy that has been donated to him. He probably chose the gpl
because it was compatible with the requirements of this letter.

>
>> Here is the letter which proves it :
>> http://maxima.sourceforge.net/misc/maxima-doe-auth.png
>
>
> Paperwork?

Yes, paperwork...

>> * Pari/GP (computer algebra system) : GPL, no trademark, no foundation
>> Was formally a proprietary project, and switched to the gpl. Here is a
>> related entry in the FAQ :
>>     "Why use such a "restrictive" license as the GPL? Will you change
>> it?
>>
>>     The short answer to the second question is no.
>>     We are quite happy with the GPL. The original license (up to versi=
on
>> 2.0) was proprietary, too casually worded, and lead to various
>> conflicts. Starting from version 2.0, we wanted a well known license f=
or
>> the whole PARI/GP package, that would leave contributors secure with t=
he
>> future use of their code and, as far as possible, not prevent anyone
>> from helping out. It was soon agreed that PARI/GP should become free
>> software.
>>     The GPL was a natural choice. It is certainly well-known, and it
>> satisfied every developer involved in the project, as well as their
>> respective employers. For the latter, the alternatives would have been
>> more restrictive, certainly not closer to the LGPL, BSD, or the Artist=
ic
>> license. Most free Computer Algebra Systems also use the GPL, presumab=
ly
>> for related reasons."
>
>
> This Henri Cohen seems to be still active in development or directly
> oversees it, probably made the license transition that much easier.

Agreed.

>> That's already a long list, although I must be forgetting important
>> ones. The important points are :
>> - some have a trademark, even individuals have (FFmpeg)
>> - some don't have, and don't seem to suffer too much (Octave, Maxima,
>> etc.)
>
>
> Trademark is for protecting the name, probably more for a business.=20
> But if some strange court case arises down the road where someone like
> Octave, Maxima, etc. loses the right to use the name, there'll be a
> mad scramble at the trademark office.

Agreed too.

>> - Interestingly enough, some projects come from proprietary products a=
nd
>> became GPLed (Pari/GP - decision explained above, Maxima )
>
>
> Pari/GP hardly seems to have originally been proprietary.=20
> Universities are slightly different than companies because of the
> where the money comes from.  (Although, universities seem to be going
> more and more down that road as even the public schools become more
> and more private because state governments lessen and lessen support.)

It's the author who said that the original license was proprietary...
"The original license (up to version 2.0) was proprietary, too casually
worded, and lead to various conflicts."

>> We can be fine without a legal organization.
>
>
> My vision is not that far, but others might disagree.
>
>> We can also define an
>> organization that supervises the development, but taking complete
>> control on releases for example would makes gnuplot look like Scilab
>> which is still sometimes criticized for its almost closed development.
>
>
> I'm thinking more a loosely legal agreement about the authority of
> maintainers.  And it probably doesn't have so much to do with the
> license.  But enough authority for stamp of approval release (no
> warranty!), but not enough to move development outside the intended pat=
h.

So we would need to define cleanly how this group will work, but if it's
properly done and functional, it's ok !

>
>> What do you conclude ?
>
>
> First, thanks again for the thorough summary.  I need some time to thin=
k.
>
> Dan

You're welcome !

Timoth=E9e

Re: Thoughts about gnuplot license and copyrights

[Daniel J S. <dan...@ie...>]

Timothée Lecomte wrote:

>>>As Hans-Bernhard Broeker also explained, a trademark would be needed to
>>>protect the name 'gnuplot'. To remain pragmatic, nobody will provide the
>>>money to have a 'gnuplot' trademark.
>>
>>
>>I'm not assuming that.  International trademark looks to be $500 (it's
>>in swiss currency) for ten years.  Not unreasonable.  But that's an
>>aside.  (Might make it easy though to recover www.gnuplot.com from
>>that person who gobbled it up.)

It's an aside.  Different but not totally unrelated issue.


>>Hence, why I propose that if a member of the group goes silent for a
>>year, someone else is allowed to take over.  If it gets to the point
>>where there isn't a dozen or so people who could be a pool for just a
>>few maintainers then nobody cares anymore.  Or do a D. Knuth thing and
>>freeze it for good.
> 
> 
> Well, as we should probably not freeze it ;-),

Twenty, thirty years from now?  The project having evolved to the point of needing little maintenance?  Who knows?

Useful list below, thanks.  Short comments within:

> I have wondered how things go for other projects. Indeed, instead of
> trying to resolve the gnuplot's license disadvantage ourselves, we can
> get the inspiration from others (that's not copyrighted !). So I have
> made a compilation of information on famous scientific open-source
> software. Let's see :
> 
> 
> * octave ( numerical computations): GPL, "octave" doesn't seem to be a
> trademark, octave is a GNU product, so it is proctected by the Free
> Software Foundation, which takes good care of copyrights (see
> http://www.gnu.org/prep/maintain/maintain.html  :  the FSF asks for a
> plain paper from each important contributor that gives his copyright to
> the FSF, as for each GNU software)

We could probably talk to John Eaton about more specifics if we wanted.  By plain paper you mean something with signatures?  (And notarized, I'd argue.)

> 
> * the R project (statistical language): GPL, no trademark, backed by the
> R foundation for general guidelines and the Free Software Foundation as
> it is a GNU project. The statutes of the R foundation are available here
> : http://www.r-project.org/foundation/Rfoundation-statutes.pdf
> 
> * Maxima (computer algebra system) : GPL, no trademark ("maxima" is
> already the trademark of a fishing lines company),

Doesn't have any bearing on a software license as fishing equipment is unrelated to software.

> no organization behind.
> Maxima comes from DOE-Macsyma, written by the department of Energy (as
> far as I understand) and was officially donated to the community under
> the GPL.

Not according to the letter.  It apparently isn't GPL but says if the derivative work is GPL then Maxima's license must also accompany the GPL license.

> Here is the letter which proves it :
> http://maxima.sourceforge.net/misc/maxima-doe-auth.png

Paperwork?

> 
> * Pari/GP (computer algebra system) : GPL, no trademark, no foundation
> Was formally a proprietary project, and switched to the gpl. Here is a
> related entry in the FAQ :
>     "Why use such a "restrictive" license as the GPL? Will you change it?
> 
>     The short answer to the second question is no.
>     We are quite happy with the GPL. The original license (up to version
> 2.0) was proprietary, too casually worded, and lead to various
> conflicts. Starting from version 2.0, we wanted a well known license for
> the whole PARI/GP package, that would leave contributors secure with the
> future use of their code and, as far as possible, not prevent anyone
> from helping out. It was soon agreed that PARI/GP should become free
> software.
>     The GPL was a natural choice. It is certainly well-known, and it
> satisfied every developer involved in the project, as well as their
> respective employers. For the latter, the alternatives would have been
> more restrictive, certainly not closer to the LGPL, BSD, or the Artistic
> license. Most free Computer Algebra Systems also use the GPL, presumably
> for related reasons."

This Henri Cohen seems to be still active in development or directly oversees it, probably made the license transition that much easier.

> 
> * Axiom (computer algebra system ) : BSD-style, no trademark ("axiom" is
> already the trademark of an automotive company),

Again, there can be multiple trademarks if the two companies are in unrelated fields.

[snip]

> * amarok ( music player ): GPL, no trademark, copyright is protected by
> the KDEev as it is part of KDE

Also the name of a Mike Oldfield album.  :-)

> That's already a long list, although I must be forgetting important
> ones. The important points are :
> - some have a trademark, even individuals have (FFmpeg)
> - some don't have, and don't seem to suffer too much (Octave, Maxima, etc.)

Trademark is for protecting the name, probably more for a business.  But if some strange court case arises down the road where someone like Octave, Maxima, etc. loses the right to use the name, there'll be a mad scramble at the trademark office.

> - some have a legal organization behind them, which has often the role
> to 'supervise' development by defining global directions (R Foundation,
> Mozilla Foundation...), mostly to verify that copyrights and trademarks
> are respected (FSF, KDEev, Mozilla Foundation, Linuxmark...), sometimes
> to control all of the development process (Scilab)
> - some had an organization, but lost it as it was not that efficient
> (wxWidgets) !
> - many are GPL or LGPL and seem to be fine with it !
> - Interestingly enough, some projects come from proprietary products and
> became GPLed (Pari/GP - decision explained above, Maxima )

Pari/GP hardly seems to have originally been proprietary.  Universities are slightly different than companies because of the where the money comes from.  (Although, universities seem to be going more and more down that road as even the public schools become more and more private because state governments lessen and lessen support.)

> - the FSF has a strong policy regarding copyrights : it asks for signed
> paper for each important contributor
> - if we want to define a legal board of maintainers, we could have ideas
> from the R foundation (see statutes above)
> - if Thomas Williams was to accept a change in the license, we can have
> an idea of the corresponding paper for Maxima (see link above)
> 
> 
> What do I conclude ?
> 
> We could be fine with the GPL. Even the linux kernel is GPLed, although
> I honestly find that it represents a lot more work as gnuplot ! So the
> GPL can protect gnuplot reliably.
> We could be fine without a trademark. It seems possible to protect the
> name 'gnuplot', but that's a different problem, and should probably be
> adressed after the license is.

Agreed.

> We can be fine without a legal organization.

My vision is not that far, but others might disagree.

> We can also define an
> organization that supervises the development, but taking complete
> control on releases for example would makes gnuplot look like Scilab
> which is still sometimes criticized for its almost closed development.

I'm thinking more a loosely legal agreement about the authority of maintainers.  And it probably doesn't have so much to do with the license.  But enough authority for stamp of approval release (no warranty!), but not enough to move development outside the intended path.

> What do you conclude ?

First, thanks again for the thorough summary.  I need some time to think.

Dan

Re: Thoughts about gnuplot license and copyrights

[Ethan A M. <merritt@u.washington.edu>]

On Saturday 21 January 2006 05:41 pm, Timoth=E9e Lecomte wrote:

> - the FSF has a strong policy regarding copyrights : it asks for signed
> paper for each important contributor

The current front page for gnuplot document lists 23 contributors
in addition to the two original copyright holders.
=20
> What do you conclude ?

I conclude that you would need on the order of 25 signatures to
fully satisfy the FSF, if that is your goal.
But that specific criterion would only be triggered if you were
attempting to transfer copyright to the FSF, which is rather extreme. =20

=2D-=20
Ethan A Merritt
Biomolecular Structure Center
University of Washington, Seattle 98195-7742

Re: Thoughts about gnuplot license and copyrights

[<tim...@en...>]

Daniel J Sebald wrote:

> Timoth=E9e Lecomte wrote:
>
>> I understand your point about defining legally this group. As Ethan
>> explained, it's what the pine license does, and in this case the group
>> which authorizes releases is the University of Washington. You can not=
e
>> that there are major differences between pine and gnuplot : 'pine' is =
a
>> trademark (as 'firefox' is for an other example), and there is a
>> powerful and long-standing legal entity behind pine, the university.
>>
>> To remain pragmatic, I think that we can't reach the same conditions f=
or
>> gnuplot.
>
>
> Why not?  I know the common belief now days is that only "big"
> entities have or can afford rights, representation, etc.  But the law
> isn't restricted to them.  In fact, a lot of these legal concepts
> probably were originally intended to protect the individual.

Ok, it is possible, but personnaly I don't want

>
>> As Hans-Bernhard Broeker also explained, a trademark would be needed t=
o
>> protect the name 'gnuplot'. To remain pragmatic, nobody will provide t=
he
>> money to have a 'gnuplot' trademark.
>
>
> I'm not assuming that.  International trademark looks to be $500 (it's
> in swiss currency) for ten years.  Not unreasonable.  But that's an
> aside.  (Might make it easy though to recover www.gnuplot.com from
> that person who gobbled it up.)
>
> Let's put it this way.  You mentioned having a little assurance that
> work doesn't get wasted.  Well, take the work of the number of people
> who have contributed and add up their time as though it were charged
> an hourly rate.  Wow, that would be one damn big monetary gift;
> intended, I think, for science.  It wouldn't kill me to contribute
> monetarily to solidifying a license so long as it's reasonable.

I completely agree with you : the work already done deserves a monetary
contribution. But solidifying the license can also be done without any
money. It's just a matter of what we want to protect, and how far we are
ready to go.

>> And if we define a "maintainers group", it would still be a list of
>> named individuals who can disappear one day.
>
>
> Hence, why I propose that if a member of the group goes silent for a
> year, someone else is allowed to take over.  If it gets to the point
> where there isn't a dozen or so people who could be a pool for just a
> few maintainers then nobody cares anymore.  Or do a D. Knuth thing and
> freeze it for good.

Well, as we should probably not freeze it ;-), I find the first proposal
a valid solution.

>
>> The solution would be a
>> sort of foundation, but I can't imagine such a complex legal entity fo=
r
>> just gnuplot.
>
>
> Not complex.  Keep it simple.  Nobody is attempting to collect royaltie=
s.

Royalties, no, but signatures, yes. I don't like paper work, but I admit
it is feasible.

>> That's why I just propose to use a slightly more permissive license
>> about release conditions. Thus we preserve simplicity : no need to
>> legally define a maintainer group if the license does not mention it.
>
>
> That's one alternative I'm saying leave on the table.
>
> I guess my feeling is that gnuplot has had some license now for a long
> time.  If out of the blue it changes to a different license it would
> be good, I think, to have it made clear that T.W. has signed off on
> that.  Certainly an email from T.W. wouldn't hold much water.=20
> (Someone could claim the email was made up by someone else.)=20
> Something he can sign (notarized) that can be put in a safe deposit
> box would be nice.  To change the license without there being a clear
> understanding of that intention would just add more confusion.
>
> Dan

Yes, if there's a change, we should be able to prove that TW has
approved it.



I have wondered how things go for other projects. Indeed, instead of
trying to resolve the gnuplot's license disadvantage ourselves, we can
get the inspiration from others (that's not copyrighted !). So I have
made a compilation of information on famous scientific open-source
software. Let's see :


* octave ( numerical computations): GPL, "octave" doesn't seem to be a
trademark, octave is a GNU product, so it is proctected by the Free
Software Foundation, which takes good care of copyrights (see
http://www.gnu.org/prep/maintain/maintain.html  :  the FSF asks for a
plain paper from each important contributor that gives his copyright to
the FSF, as for each GNU software)

* the R project (statistical language): GPL, no trademark, backed by the
R foundation for general guidelines and the Free Software Foundation as
it is a GNU project. The statutes of the R foundation are available here
: http://www.r-project.org/foundation/Rfoundation-statutes.pdf

* Maxima (computer algebra system) : GPL, no trademark ("maxima" is
already the trademark of a fishing lines company), no organization behind.
Maxima comes from DOE-Macsyma, written by the department of Energy (as
far as I understand) and was officially donated to the community under
the GPL. Here is the letter which proves it :
http://maxima.sourceforge.net/misc/maxima-doe-auth.png

* Pari/GP (computer algebra system) : GPL, no trademark, no foundation
Was formally a proprietary project, and switched to the gpl. Here is a
related entry in the FAQ :
    "Why use such a "restrictive" license as the GPL? Will you change it?

    The short answer to the second question is no.
    We are quite happy with the GPL. The original license (up to version
2.0) was proprietary, too casually worded, and lead to various
conflicts. Starting from version 2.0, we wanted a well known license for
the whole PARI/GP package, that would leave contributors secure with the
future use of their code and, as far as possible, not prevent anyone
from helping out. It was soon agreed that PARI/GP should become free
software.
    The GPL was a natural choice. It is certainly well-known, and it
satisfied every developer involved in the project, as well as their
respective employers. For the latter, the alternatives would have been
more restrictive, certainly not closer to the LGPL, BSD, or the Artistic
license. Most free Computer Algebra Systems also use the GPL, presumably
for related reasons."

* Axiom (computer algebra system ) : BSD-style, no trademark ("axiom" is
already the trademark of an automotive company), no foundation but the
devs seem to think of it (see
http://wiki.axiom-developer.org/AxiomFoundation , its role would be to
promote axiom development).
Axoim was an IBM software, then sold to "NAG", which stoped to sell it
(not a financial success), and released it under a BSD-style license.

* Ginac ( "ginac is not a computer algebra system") : GPL, no trademark,
no organization

* Scilab ( numerical computations ) : open source custom license
described as non-free by the FSF, as it does not allow commercial
distribution of the modified code. "Scilab" is a trademark, the Scilab
Consortium controls its development.

* Jacal (computer algebra system ) : GPL, no trademark, no organization

* Boinc (distributed computing ) : LGPL, no trademark, copyright holder
is the University of California


And now other more general software :

* the linux kernel : GPL, "linux" is a tradermark, a non-profit
corporation is there to protect the trademark
(http://www.linuxmark.org/), the rest is self-organized around Linus
Torvalds

* Gnome : LGPL (can be linked to proprietary code), "gnome" is a
tradermark, the Gnome foundation is behind it

* KDE : GPL (as Qt), "kde" is a tradermark, the KDE.e.v is a non-profit
organization behind it (http://ev.kde.org/)

* Xfce ( desktop environment ): GPL/LGPL/BSD depending on files, no
trademark, no foundation

* Xorg : original X11 license, very permissive, non-copyleft (ie allow
to impose further restrictions on a distributed code), "xorg" doesn't
seem to be a trademark, there is limited-liability-company, the Xorg
foundation, that supervises the development

* XFree86 : modified X11 license (gpl incompatible, that causes the
fork), "xfree86" is a trademark, XFree86 Project Inc. is a company behind=
 it

* firefox, thunderbird : MPL (gpl incompatible due to a few
restrictions), "firefox", "thunderbird", "mozilla" are trademarks,
protected by the Mozilla Foundation, a non-profit organization

* wxWidgets : modified LGPL (more permissive), "wxWidgets" is not a
trademark (they already changed from wxWindows due to trademark problems
with Microsoft), there used to be a "wxWindows foundation", but it
disappeared due to the lack of consensus and fears about copyrights that
were given to the foundation (I quote http://www.wxwidgets.org/sf/ :
"However, assignment was perceived by some to have serious
disadvantages, such as the problem of what to do with the IP should the
Foundation be dissolved." )
=3D=3D>That's interesting !

* FFmpeg ( audio/video toolkit ) : LGPL (+ GPL modules), "FFmpeg" is a
trademark of the author of the project, no organization

* amarok ( music player ): GPL, no trademark, copyright is protected by
the KDEev as it is part of KDE

* Stellarium ( planetarium ) : GPL, no trademark, no organization

* Celestia ( space simulation ) : GPL, no trademark, no organization

* Inkscape ( svg graphics ) : GPL, no trademark, no organization



That's already a long list, although I must be forgetting important
ones. The important points are :
- some have a trademark, even individuals have (FFmpeg)
- some don't have, and don't seem to suffer too much (Octave, Maxima, etc=
.)
- some have a legal organization behind them, which has often the role
to 'supervise' development by defining global directions (R Foundation,
Mozilla Foundation...), mostly to verify that copyrights and trademarks
are respected (FSF, KDEev, Mozilla Foundation, Linuxmark...), sometimes
to control all of the development process (Scilab)
- some had an organization, but lost it as it was not that efficient
(wxWidgets) !
- many are GPL or LGPL and seem to be fine with it !
- Interestingly enough, some projects come from proprietary products and
became GPLed (Pari/GP - decision explained above, Maxima )
- the FSF has a strong policy regarding copyrights : it asks for signed
paper for each important contributor
- if we want to define a legal board of maintainers, we could have ideas
from the R foundation (see statutes above)
- if Thomas Williams was to accept a change in the license, we can have
an idea of the corresponding paper for Maxima (see link above)


What do I conclude ?

We could be fine with the GPL. Even the linux kernel is GPLed, although
I honestly find that it represents a lot more work as gnuplot ! So the
GPL can protect gnuplot reliably.
We could be fine without a trademark. It seems possible to protect the
name 'gnuplot', but that's a different problem, and should probably be
adressed after the license is.
We can be fine without a legal organization. We can also define an
organization that supervises the development, but taking complete
control on releases for example would makes gnuplot look like Scilab
which is still sometimes criticized for its almost closed development.


What do you conclude ?


Timoth=E9e

Re: Thoughts about gnuplot license and copyrights

[Daniel J S. <dan...@ie...>]

Timoth=E9e Lecomte wrote:

> I understand your point about defining legally this group. As Ethan
> explained, it's what the pine license does, and in this case the group
> which authorizes releases is the University of Washington. You can note
> that there are major differences between pine and gnuplot : 'pine' is a
> trademark (as 'firefox' is for an other example), and there is a
> powerful and long-standing legal entity behind pine, the university.
>=20
> To remain pragmatic, I think that we can't reach the same conditions fo=
r
> gnuplot.

Why not?  I know the common belief now days is that only "big" entities h=
ave or can afford rights, representation, etc.  But the law isn't restric=
ted to them.  In fact, a lot of these legal concepts probably were origin=
ally intended to protect the individual.

> As Hans-Bernhard Broeker also explained, a trademark would be needed to
> protect the name 'gnuplot'. To remain pragmatic, nobody will provide th=
e
> money to have a 'gnuplot' trademark.

I'm not assuming that.  International trademark looks to be $500 (it's in=
 swiss currency) for ten years.  Not unreasonable.  But that's an aside. =
 (Might make it easy though to recover www.gnuplot.com from that person w=
ho gobbled it up.)

Let's put it this way.  You mentioned having a little assurance that work=
 doesn't get wasted.  Well, take the work of the number of people who hav=
e contributed and add up their time as though it were charged an hourly r=
ate.  Wow, that would be one damn big monetary gift; intended, I think, f=
or science.  It wouldn't kill me to contribute monetarily to solidifying =
a license so long as it's reasonable.

> And if we define a "maintainers group", it would still be a list of
> named individuals who can disappear one day.

Hence, why I propose that if a member of the group goes silent for a year=
, someone else is allowed to take over.  If it gets to the point where th=
ere isn't a dozen or so people who could be a pool for just a few maintai=
ners then nobody cares anymore.  Or do a D. Knuth thing and freeze it for=
 good.

> The solution would be a
> sort of foundation, but I can't imagine such a complex legal entity for
> just gnuplot.

Not complex.  Keep it simple.  Nobody is attempting to collect royalties.

>=20
> That's why I just propose to use a slightly more permissive license
> about release conditions. Thus we preserve simplicity : no need to
> legally define a maintainer group if the license does not mention it.

That's one alternative I'm saying leave on the table.

I guess my feeling is that gnuplot has had some license now for a long ti=
me.  If out of the blue it changes to a different license it would be goo=
d, I think, to have it made clear that T.W. has signed off on that.  Cert=
ainly an email from T.W. wouldn't hold much water.  (Someone could claim =
the email was made up by someone else.)  Something he can sign (notarized=
) that can be put in a safe deposit box would be nice.  To change the lic=
ense without there being a clear understanding of that intention would ju=
st add more confusion.

Dan