[GRASS5] v.in.dwg license problem

Bernhard Reiter bernhard at intevation.de
Wed May 14 07:14:05 EDT 2003


Radim,

first please be reassured that I'm as unhappy as you about the situation, 
but I'm thinking about the long term availability of GRASS as Free Software.
Thus we need to clearly address the issue.


On Wed, May 14, 2003 at 11:50:52AM +0200, Radim Blazek wrote:
> On Tuesday 13 May 2003 07:25 pm, Bernhard Reiter wrote:
> > Noticed that v.in.dwg from GRASS 5.1
> > (http://freegis.org/cgi-bin/viewcvs.cgi/grass51/vector/v.in.dwg/)
> > uses the proprietory library opendwg.
> > As I believe that it also needs the GRASS libraries which
> > are under GNU GPL, this means that v.in.dwg has a severe license problem.
> 
> Why? Which paragraph of GPL exactly is violated by v.in.dwg SOURCE
> CODE distribution.

There are two problems with using opendwg:
A legal one and a strategic one.
I've touched both briefly in my last email.

The legal one:
It is tricky as licenses can only have an effect when software is distributed.
Linking to the libraries under GNU GPL with opendwg 
and distributing the binaries is a violation of the GNU GPL.
The person doing this looses the right to use or distribute that library.
The license requirements of v.in.dwg basically calls for that violation.
The statements of the GRASS team assume that GRASS is under GNU GPL
without introducing these kind problems. 

So we need to remove that code from the main GRASS CVS.
We can keep it at another place.

On the strategic one, compare to
http://www.fsf.org/licenses/gpl-faq.html#WritingFSWithNFLibs

> > I see no viable alternative than to remove the current v.in.dwg
> > and work on a replacement based on a Free Software license
> > (compatable with GRASS' license).
> > DIME or dxflib might be a starting point for development.
> 
> OpenDWG was not choosen by chance. 

I know, but it has the main drawback to not be Free Software.
If there is not appropriate library available we must start
to write a better one.

> DIME: C++ (not welcome in GRASS)
>       DXF only (no DWG)
>       versions 10-14 guaranteed only (missing 2000 and 2002)
> dxflib: C++
>         DXF only 
>         "It's at the moment very simple ..." (from home page)
> 
> > We should remove v.in.dwg because nobody can distribute binaries
> > and if somebody did, this person would violate the license of GRASS
> > which mean he strictly would loose the right to use GRASS.
> 
> What is the problem? Just do not ditribute v.in.dwg binaries!

That means v.in.dwg is not Free Software and does not inherit
the other advantages GRASS has.

> OpenDWG library is not distributed with GRASS. Everybody who
> want to compile v.in.dwg must download OpenDWG library and 
> explicitly enable v.in.dwg compilation, so he knows about OpenDWG
> license and cannot distribute v.in.dwg binaries just by accident.

We can place it at a different place,
to keep the usefulness of the application.

> As copyright holder, I will cover v.in.dwg by GPL with exception 
> to link it to proprietary libraries. Once we have some 
> more reasonable license for GRASS libraries, this will allow
> to sell v.in.dwg in binary form for those having proper license 
> from OpenDWG (commercial associate membership etc.)

Strategically that is not a good idea.

> > It might hurt in the moment, especially if v.in.dxf works nicely,
> 
> What do you mean? v.in.dxf DOES NOT work nicely. 

That was a typo, I appologise.
I meant "v.in.dwg" which I was talking about the whole time.

> It supports 
> DXF only up to version 9 or 10! If you realy want to exchange
> data with ACAD users (there are many) you need some proprietary SW 
> (or OpenDWG convertor for Windows ;), to convert DWG/DXF files
> you have got, to some very old DXF version to enable v.in.dxf import.

Proprietory software is unacceptable in the long run.

> > but if we don't pay attention to GRASS' freedom, we'll
> > pay a enourmously higher bill in the long run hurting us plenty.
> 
> Can you concretize these bills?

This would mean explaining the advantages and non-advantages
of proprietory software. Probably a bit beyond the scope of this thread. 
David Wheeler as assembled nice evidence on the technical points.
http://dwheeler.com/oss_fs_why.html
The FSF has assembled a lot of arguments for the ethical reaons.
E.G. http://www.fsf.org/philosophy/why-free.html

For GRASS is would mean the problems that have been there before:
Cannot be taken up easily by major distributors, thus in general
GRASS would not have that capabilities. Dependency on the
proprietory software producer and so on.

> > Another possiblity apart from using DIME, dxflib or develop
> > a new library would be to lobby
> > the opendwg consortium http://www.opendwg.org/ to release
> > opendwg under a reasonable Free Software license (X11 Style or GNU
> > LGP come to mind). To do this successfully is probably hard.
> >
> > I don't consider relicensing GRASS' core libraries under a more relaxing
> > license an option. 
> 
> That is more general question. 

I'll answer that in another email.
	Bernhard
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://lists.osgeo.org/pipermail/grass-dev/attachments/20030514/e8c5f37d/attachment.bin


More information about the grass-dev mailing list