[Live-demo] OSGEO Live GIS , ideas to share

Alex Mandel tech_dev at wildintellect.com
Sun Dec 13 12:09:09 PST 2009

Massimo Di Stefano wrote:
> Hi All
> some ideas about the osgeo's live gis project, i'd like to know your opinion about :
> - can we fix a date to start the up-to date work needed for the install scripts to runs on unbuntu 9.10 ?
> - what about adding different architecture e.g. to have install scripts running on "i386, x86_64, ppc" ?
> - can we contact the local chapter's asking for a "tester",  so we can try to adapt the scripts to run on different localization ?
>   |--  what do you think to add a directory in the svn tree : "localization" and subdir for different language/local-chapters ... us, en, fr, it, de, ... ?
> - linux and geoscience .. i'd like the idea to have on a "dedicated" osgeo live project, not only speciphic "geo" sw, i agree the we are on Osgeo, but adding other science sw (like plotting tools, statistic and symbolic calculus sw .. etc ... ) can be a cool start point for an live project dedicated to an "educational" release  (in this case ,"educational-release", i vote to eliminate the mac and win installer and give more focus (and space) on new sw, tutorial data and documentation)
> thanks,
> Massimo.

My previous comments were about priority not about end goal content. I
do think that including things like R, GGOBI, etc is a great add to the
disc especially since it integrates with other geospatial software. I
would turn down things like Octave that are really unrelated to
geospatial though.

What I intended to push was just that we focus on getting OSGeo projects
onto this next release. That way for future releases we can just update
them and shift our focus to selecting the best "other" software to include.

For the disc that we hand out the windows/mac installers to me are an
important component. With all live discs I've used in the past, not all
people are ready to jump to using linux, and having those installers
handy for those that want to move from using the live to installation
allows them to join the community quickly. I actually helped someone
install QGIS/GRASS onto their Mac at the last conference I was at,
thanks to the live disc I had.

I hesitate to get into the architecture issue now too. We defaulted to
i386 in order to ensure that the disc would run on the largest possible
number of machines and reduce our complexity in building. Ideally as we
move everything to deb packages and build them against multiple
architectures this would solve itself and a metapackage would suffice
that would be available on all the architectures. I also think we should
postpone working to much on this until a future release do to complexity
that it adds.

One way to think about the concept is that we are putting together the
geospatial live disc/vm/packages, which we would gladly contribute to
Edubuntu, Posiedon etc for inclusion in their core. But we should limit
ourselves to focusing on geospatial in order keep this task manageable,
maximize our use of resources(people time), and not be redundant with
all the other great working in FOSS out there.

Keep in mind this is a base for others to build on, if you have time to
take the core here and build 64 bit with the extra software you want
your audience to have, that's awesome. But being realistic for the rest
of us, we need to keep the scope manageable with the resources we have.


More information about the Osgeolive mailing list