[OSGeo-Discuss] Use cases for FOSS-GIS in universities

[Ari Jolma]

Tim Michelsen kirjoitti:
> Hello,
> I don't know if this is the right place to discuss this but I just start.

We have OSGeo-edu, a working group and a list. The edu-list might be a 
better place to discuss this, so I include all your message in this and 
set replies to edu_discuss at lists.osgeo.org.

>
> At my university open source is little used in eductaion and research. 
> There are only a few single warriors who care about it.
>
> We have serveral departments that use GIS programs. Each of them buys 
> normally the licences for Arc*
>
> Most students make their way into one program after attening a number 
> of classes on how to push the buttons of that program. But when they 
> come back to it after a while all this is lost since they haven't 
> learned the logic behind. To my thinking, getting into the many FOSS 
> programs forces students to lokk behind and learn concepts rather than 
> functions.

I suspect everybody thinks and says this but then go on and produce 
exercises that are basically pushing the buttons or copy-paste. I often 
use the latter approach. My experience is that unless you use a single 
program from the beginning to end through many courses, it's not 
feasible to expect that the students know the program/add-on well enough 
so they can focus on the concepts rather than specifics of the program. 
At least in the exercise itself where time is limiting. And when you 
start arguing for _one program_ over many courses, it is very easy to 
speak in favor of "the industry leader", which "is used everywhere" and 
with which "students get marketable skills".

>
> Another thing is the question of resources. Instead of buying licences 
> from big companies that money could be saveed to by data loggers, 
> equipment or pay a FOSS-developer.
>
> Most projects buy expensive tools when they only want to produce some 
> maps to display the survey fields etc.
>
> So, my question is:
> * Is there a possibilty set up a university wide infrastructure on 
> FOSS that enables whoever neeeds it to handle geodata and analyse it 
> even when they are not educated GIS specialists (rather geo/agric 
> scientists)?

I guess this can be done, but at least in our case different departments 
and students/faculty in them are very different, at least in their 
attitude towards software. A general rule seems to be that CompSci 
students embrace easily Linux but NatResMgmt students embrace Windows. 
Of course there are exceptions, but departments, even units are quite 
independent in their decisions and in an individualistic and free 
environment like university it is impossible to dictate what any faculty 
uses in their research/teaching. You can try of course and I wish you 
good luck.

>     * Naive idea/vision:
>         1) computing center of the university employs a GIS 
> specialist(s) who act as service force for other disciplines (set up 
> of geodatabases, introductory courses)
>         2) computing center sets up a server with GRASS, postgis, etc.
>         3) those who need geo processing will install a tailored cywin 
> or eny other environment to access the latest version of the FOSS GIS 
> software on the server via -X forwarding or simply access their data 
> in the postgressdb from various clients.
>         4) data in the postgressdb could be shared according to given 
> access rights
> => there infrastruture is just there, those who need take to whatever 
> level they'd need it.
>
> * Question: would it be possible to implement such a scenario?
>
> * Are there already such cases out there?
>
> * Why not take the nice example of the various projects that deliver 
> FOSS for schools (Edubuntu, Skolelinux, etc.) and adapt this to the 
> world of FOSS4G?
>     
> * Who or what are the thoughts of OSGEO on this?
>
> Keen to hear your opinion, thoughts, experineces, critics, etc.

I've been using quite a bit a stack that I've built myself 
(Geoinformatica, 
http://geoinformatics.tkk.fi/twiki/bin/view/Main/GeoinformaticaSoftware). 
My experiences are a bit mixed, I think especially the programming part 
(mostly copy paste Perl code into a text input box in a GUI) is very 
challenging for all my students (I've required that only from Master's 
students but these are often civil engineers who are not required to 
learn programming at all). However, I like very much to use it since I 
know it completely (which I can't say of the proprietary products that 
we have available), and I believe that at least _I_ can focus on the 
concepts rather than on the interface in my lecturing/leading an exercise.

I've been forced to port the software to Windows and it is now quite 
easy to use in our University's Windows class system since it does not 
require a specific installation, it runs from any directory. However, 
since PostGIS data storage requires installation, I've not yet used it, 
but probably need to do that for the next semesters. Similarly, I'm just 
only getting to start using Mapserver in my teaching. With these server 
products (PostGIS and Mapserver) the requirement is that we and 
preferably also the students have admin rights, and in practise that 
means our own systems. I'm not the only professor in our unit and I 
don't have much command over the scarce resources that we have, so I've 
so far not used them much. I've tried setting up a Linux box of myself 
and letting the students log on to that over the network and use X (the 
benefit is that you can do that from Windows or X terminals) but there 
are problems with that approach.

Although my University has a large Unix system also and while it runs 
many debian-based systems, my experience is that it is even more 
difficult to set up a free GIS environment (the Geoinformatica for 
example) there - so in fact the Windows solution is not so bad after 
all. And as students usually have Windows machines at home, at least 
some students have been succesful in downloading the system , installing 
it, and using it. We have some aspirations for remote education too and 
in such a setting a freely downloadable system is indispensable.

Ari

>
> Tim
>
> _______________________________________________
> Discuss mailing list
> Discuss at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/discuss