[Freegis-list] Impressions from OSG '05

[Jan-Oliver Wagner]

Hi FreeGIS community,

finally I returned to my office after staying in Minnesota
OSG '05 and right after being at Linuxtag 2005 in Karlsruhe,
Germany.

I enjoyed OSG '05 very much having met some people in real
life for the first time. Unfortunatly I was not able to
talk to all the guys I liked to talk with nor to attend
all the interesting presentations, but here are at least
some personal impressions:

Day 0:
At Stubs&Herbs some of us had a good time meeting the
first time and having some discussion over a beer.
For the first time I heard of the company Orkney which
is located in Japan and boxes GRASS and MapServer in
japanese. This all look really professional (despite the
fact that I do not speak japanese ;-) and Toru Mori
gave me some interesting insights into Free GIS Software
market in Japan.

Day 1:
The PostGIS workshop held by Paul Ramsey
was just great. Paul was always right to the point and
I recommend to attend this workshop whenever you have
a chance in the future.
I hope Paul will prepare an extended version of this
workshop which IMHO would be very beneficial since
PostGIS really rocks.

In the afternoon I attended the Chameleon workshop.
Chameleon is a framework for building user
interfaces for web mapping applications. This workshop
was more directed towards developers and what I learned
is that Chameleon is not yet easy to use but very
powerful once you have managed to implement your
first Chameleon widget (which we did during this
workshop :-). The workshop was given by Paul Spencer.

What I learned from both workshops is that I am not
used to be productive on Windows systems ;-)
(Where is tab completion and why is copy&paste so complicated?)

At night we had a typical-US BBQ with lots of Beer again
at Fort Snelling where we have been taught the gunners
job. Paul (Chameleon) did this job well, but I still don't
know which target he was aiming at :-)

Day 2:
The plenary session consisted of a number of lightning
talks and the featured talk of Markus Neteler about
GRASS.

Of the lightning talks the visually most impressive
one was Norman Vine showing his osgPlanet. This is
hot stuff and I hope so much it will run on GNU/Linux
eventually.

Sean Gillies explains his latest work and ideas on
Python-impemented GIS stuff. I think he is right in that
it is more important right now to bring GIS to the Zope community
rathern than vice versa.

Schyler Erle dropped his topic on his new book "Mapping Hacks"
in favor of a very enthusiastic call for action on a
sort of distributed WMS caching with peer-to-peer elements
to lower the load on the original WMS servers which partly
got switched off due to too high load.
(If I got it right).

Markus Netelers talk about GRASS seemed to impress most of
the crowd. GRASS 6 has full vector support and the
raster part is better than ever. I guess the downloads increased
after that talk a lot ;-)
However, GRASS 6 has been done by only a little, very engaged,
core group. Imagine how GRASS would look like if some more
developers could be attracted (yes, I mean _you_ :-).

The rest of the day I attended the business track which IMHO
did not came to the depth I hoped.
Notable was that Ionic representative explained that
"using Free Software needs additional staff resources".
He meant the additional work to discuss things with the other
external developers and submit contributions.
I do not agree with this statement as I think it
is the other way around: how big would be the staff
if a solution can not be based on a Free
Software product but rather must be completely
implemented anew?

The IMHO best presentation of this track was given
by Dave McIlhagga about DM Solutions. He made clear
that DM Solutions follows (increasingly strictly)
a Free Software business model and proves to be very
successfull with it.


Visiting the booth of Camp2Camp I learned that
they are having nice success on the french-speaking market
in france and switzerland. Last time I met Daniel
Faivre (at Libre Software Meeting in Metz 2 years ago)
they were only about 5 people, now they are upto 15.


Also, I talked to Allan Doyle about EOGEO.
This is something that deserves support.
I regretted to not have attended the EOGEO workshop
at Day 1.


Day 3:

A 'must' for me was the presentation of Jo Walsh
about openstreetmap.org. I think the way Jo and her
team approaches the topic of gathering free vector
roadmaps is most promising one I have seen in the past.
Also it was an exceptional style of the presentation
compared to the rest (it appeared to be inspired
by Tufte's Visual Revelations :-).

The invited talk for closing session was given
by Dirk-Willem van Gulik of Apache Software Foundation.
He explained who ASF works and how they handle the software
development focusing on quality assurance processes.
In my opinion, and as Dirk also underlined, it is not
reasonable to copy the whole ASF concept and apply
it to UMN MapServer. But his talk already inspired
to set up some improved processes for code quality
as you can read recently on the MapServer developer
mailing list.

Sol Katz Award:
When starting FreeGIS I used Sol's archive and it
helped a lot. Unfortunately I never was in contact
with him.
I was a great pleasure to see Frank Warmerdam receive
this award for the work on many base GIS libraries we
find in use in virtually any major Free GIS tool.

It followed podium discussions of which I choose
the one on 'Making Free Software Our Business'.
Again, here they did not get to the point and I
disagreed with a number of statements.

As a side note, a representative of AutoDesk said that they
will start supporting Free Software and its
community.
Whoever was in doubt of that was proven wrong later
at Sally's where AutoDesk took over the beer bill
for a table of Free Software developers :-)
More seriously, I think AutoDesk (as other proprietary GIS
companies) will slowly approach the phenonemon (in their sense) "Free
Software". Maybe we see next year a first small (unimportant)
library of AutoDesk release as Free Software as an
experiment. I only hope that they do choose a standard
license and to not tell their lawyers to yet write
another one. Their experiment will likely
be a failure if they do so.



Thanks again to Steve Lime and the whole organizing team
for this event!

Best

	Jan
-- 
Jan-Oliver Wagner               http://intevation.de/~jan/
Intevation GmbH                      http://intevation.de/
Kolab Konsortium               http://kolab-konsortium.de/
FreeGIS                                http://freegis.org/