Ominiverdi's blog - steko

Bolsena OSGeo Hacking Event

nospam@example.com (steko) — Thu, 19 Jun 2008 10:35:12 -0400

On last Monday, I took a day off to go to Bolsena and take part in the much awaited OSGeo Hacking Event that was organized by Jeroen Ticheler. As I always repeat, I'm not a developer so I had little interest in the hacking part of the event: I went there to meet people and talk with them, wearing the famous orange t-shirt.

Let alone the wonderful venue and the warm, sunny weather we were waiting from so long time... it has been a really interesting day for me. I've had a chance to talk with so many people: Jeroen Ticheler, Arnulf Christl, Simone Giannecchini, Martin Landa, Lorenzo Becchi, Andrea Cappugi, Andrea Aime and many others.

I've done my best explaining to everyone what GFOSS.it is and what we're doing (now and in our future projects). And I've also tried to understand what is going on in other countries (i.e. Germany, Czech Republic, Spain, Netherlands) and what do people at OSGeo think about the Italian Local Chapter. GFOSS.it is one of the largest and more active local chapters and our activity is well known and much appreciated, even though we're not only devoted to free open source geospatial software (OpenStreetMap being the other most followed activity). So it's a pity that not everyone from Italy is involved in the "local chapter"/GFOSS.it activity (just think about the OSGeo Charter Members: only 2 / 4 are GFOSS.it members). But we're really doing our best to involve everybody, so it's just a matter of time IMHO.

I enjoyed meeting and talking with so many people from abroad even though in the beginning I was astonished when I realized that no one from GFOSS.it was going to Bolsena. There is clearly a lack of communication between our national community and the OSGeo international community, in both directions. Everyone involved should do her/his best to “build bridges” across national borders and seas!

OSGeo Charter Member Election 2008

nospam@example.com (steko) — Wed, 11 Jun 2008 03:42:20 -0400

The results of the OSGeo Charter Member Elections 2008 are available!

Another Italian is in: Paolo Cavallini, president of the GFOSS.it association, which is also the OSGeo Local Chapter for Italy. This means there are 4 Italians among the charter members right now, even though ominoverde is living abroad. At least now we can gain some better reputation than with Euro2008...

YAFUG (Yet Another Fucking Useless GUI)

nospam@example.com (steko) — Sat, 03 May 2008 15:04:18 -0400

If you're into GFOSS development, you have to check out this rant by Francesco P. Lovergine (frankie): Che male abbiamo fatto? (scherzando ma non troppo).

The e-mail is in Italian but it shouldn't be too difficult to catch the overall message: less development of new GUIs, more stability and performance for the existing ones (that look quite bad anyway).

I agree with him, definitely. Adding tons of new features is an exciting task for a developer, but if this means more and more bugs and less usability, well, I'm better off with plain old command line.

Arthur C. Clarke 1917-2008

nospam@example.com (steko) — Wed, 19 Mar 2008 05:35:24 -0400

Arthur C. Clarke 1917-2008.

And now what is our future going to be without him?

Using the OGR Python bindings for the study of ancient monuments

nospam@example.com (steko) — Wed, 27 Feb 2008 05:45:03 -0500

In the last few months I've been experimenting a bit with the OGR Python bindings, but for a true research project. At ISCUM, the architectural study of ancient and medieval monuments has a long tradition, and though I'm not an ISCUM member, I collaborate with them since 2003.

The research topic is: statistical analysis of medieval stone walls. We measure at least width and height for each stone to be able to do some basic exploratory analysis.

Until now, the preferred method for recording stone measures was by hand. This is not so strange as it might sound: historical buildings are often in a bad conservation state, so a human operator is needed in most cases to decide which are the significant measures to be taken.

Starting from a vector-digitized image of the wall, with each stone represented by a polygon, our experiment was in first place to find algorhythms that can give similar results to a human operator when calculating these measures. Then we needed a way to divide those stones into courses (rows), to be also able to perform course-wise statistical analysis.

We did it using Python, the OGR library and R through the RPy library. Measures are calculated with Python functions that take coordinate lists as arguments. Just to give you an idea, the simplest (and inefficient) method for calculating height is like this:

 def simpleAlgo(listey):

     '''basic algo for height'''

     yMin = min(listey)

     yMax = max(listey)

     yD = yMax - yMin

     return yD

This list of Y coordinates is extracted from the OGRGeometry POLYGON object, with an intermediate step through the boundary LINESTRING.

Our base data is in a OGR-readable file (GML, Shapefile..). Internally, all data is copied into a Memory OGRDataSource so we can add fields for storing into attributes the calculated measures. Finally, the resulting OGRLayer can be saved into a new OGRDataSource like another Shapefile, a GML file, or even a PostGIS database.

Graphs are made with RPy for now. We are considering also matplotlib as a good alternative, even though R is used also in the "course finder" algorhythm.

All made with free software, all free software (under the GNU GPLv3): here is the mercurial repository that contains our code and some sample data. Contributions are more than welcome.