[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: [DebianGIS] cfitsio3 goes GPL in Debian, so what about GDAL?



On Sun, May 13, 2007 at 10:00:43PM +1200, Hamish wrote:
> Francesco P. Lovergine wrote:
> > About licensing I have to note also that we have some pending issue
> > that I hope strongly will be solved before lenny release, else we will
> > have to split gdal in a non-free and free components or in the worst
> > case move it into non-free. The latter would imply that a great part
> > of d-gis would move to contrib (i.e. out of Debian).
> 
> Would it be useful to maintain an "issues" page on the wiki for these?
> Maybe more reliable than trusting our memories + searching the ML
> archives.
> 
> Perhaps we could make use of the GDAL plugin infrastructure for
> GPL-problematic GDAL drivers? Right now AFAIK the grass-plugin is the
> only thing that uses that (libgdal-grass source pkg); is there any
> reason why other drivers couldn't be split out as well? (and then we
> just keep those drivers in the Debian/contrib archives for the legally
> adventurous).
> 

I don't know, the current main issue is in EPSG data (-> geotiff at least), 
and a few other possible issue are pointed in GDAL sources (see the PROVENANCE file).

> 
> 
> Moving off on a tangent- currently in Unstable we have binary packages
> named libgdal1-1.4.0, but version 1.4.1. From gdal.org NEWS file: "The
> GDAL 1.4.1 release is a stable release, and contains the following bug
> fixes. There should be no noteworthy changes in interfaces or APIs."
> 
> Would it be better to name those packages libgdal1-1.4, or are we just
> lucky this time that the changes are minimal enough not to warrant a
> change. Or has the ABI changed and the packages should now be in new
> ones named libgdal1-1.4.1*?
> 

I will retain the names until required for an API change, to avoid 
requeueing in NEW and abuse ftpmasters patience. I see no reasons 
to changes names. Also there is the possibility that a future
1.4.x version would break the C++ API as in the past, so I would
avoid to be too confident in a different versioning scheme.

> 
> 
> And onto another tangent (while we are peripherally on the subject of
> Debian approved licenses), I hope it will be possible in future to
> package the GRASS sample data datasets for Debian/main.
> 

It's much more interesting for me packaging the grass add-ons. 
The spearfish data set can be easily downloaded by known sources
and probably a licensing screening should also be performed on those
data. Feel free to start that analysis before packaging.

> These are the venerable Spearfish County, South Dakota, dataset and the
> soon-to-be-released North Carolina dataset* that Helena Mitasova and
> Markus Neteler are preparing.
> [*] http://article.gmane.org/gmane.comp.gis.grass.user/18796
> 
> AFAIK these will be useful as QGIS sample data as well, as long as the
> qgis-plugin-grass package is installed.
> 
> Has there been any resolution to The Great Debian Documentation License
> Debate, and if so how does this affect geo-data? ie what is an
> acceptable data license and what debian-legal post-debate guidance can
> we provide to Helena & Markus as they get ready to release?
> 
> 
> Proposed package names: "grass-data-spearfish" is easy; the
> grass-data-nc [county name?] would probably need to be split into
> several packages as, e.g., the satellite imagery may be large.
> 
> FYI, other FOSS community & Free geodata projects:
>   http://publicgeodata.org  &  http://www.osgeo.org/geodata
> 
> 
> thanks,
> Hamish

-- 
Francesco P. Lovergine



Reply to: