Saturday 21 March, late evening: Hackfest Day 3

Ok not much blogging today - but we put up some spiffy videos of ourselves.

Juergen and Martin have been pouring over the Postgres provider code to try to improve the performace.

Martin has been working on the new table integration.

Carson QgsGeometry and the QGIS Analysis library, writing unit tests and porting java analysis stuff to C++.

Werner has got the German translation of Stable branch to 100%. He also did his first steps in C++ to fix a bug in the display of the About box on non Mac platforms.

Otto has been working on making HTML version of the QGIS User Manual and updating the German edition of the user manual programmers guide.

Borys has been bringing more eyecandy to QGIS - implementing proper theme support for all core plugins. He also has QGIS Polish translations 100% complete in stable branch.

Tim has patched in WMS search capability to the WMS selector using the carto open search WMS engine. Also been cleaning up some UI issues.

Number of pizzas consumed: 5 and counting....

Friday 20 March, Afternoon through to late evening: Hackfest Day 2

Martins new table implementation has gone into SVN. The new table uses Qt model/view paradigm and should offer considerable improvements in performance for large tables.

Borys has committed improvements to the Plugin Installer so that you can choose automatic update interval and whether you only want to see experimental plugins / stable plugins etc.

Carson added a simplify feature method to QgsGeometry and implemented an accompanying unit test. All part of our nefarious plot to turn Carson into a C++ core contributor.

Juergen fixed bugs in GRASS support.

Werner and Juergen implemented automatic reporting of translation status via a nifty bash script.

Otto got busy splitting the manual into discrete parts to separate the users guide from the coding and compilation guide.

I stubbed in a unit test framework for the new QgsVectorAnalyzer class.

Added a few more piccies to my flickr account here:

Werner Macho

Martin Dobias

From left to right, Marco Hugentobler, Carson Farmer and Juergen Fischer

Andreas Neumann

Otto Dassau (left) and Werner Macho (right)

Borys Jurgiel (left) Marco Hugentobler (right) Carson Farmer (background)

Friday 20 March, Morning: Hackfest Day 2

Martin Dobias arrived! I spent more time trying to make QGIS Windows MSVC build run on a clean environment with no OSGeo4W or MSVC present. Its running but python is failing to load at launch time....still investigating why...

Carson has been working on porting his analysis functions to C++ and into libqgis_analysis.

Martin gave us a run down on his QGIS group's out-of-SVN activities:

- He showed us the cool new rendering capabilities of his symbology
- They are working on an OSM data provider and editor
- They are working on enhancment to the editing tools so that you can do things like: select mutiple vertices, move the selected vertices, remove selected vertices, undo and redo edits and so on.

Marco gave us a run down of the vector overlay branch. The vector overlay allows the intelligent placement of symbols and charts overlay features.

Otto Dassau also chatted a bit to me about his plans for documentation. They will follow the unstable QGIS releases with unstable doc releases which will be marked as such.

Thursday 19 March, Afternoon: Hackfest Day 1

Present:

Borys Jurgiel
Marco Hugentobler
Otto Dassau
Horst Duester
Carson Farmer
Werner Macho
Juergen Fischer
Tim Sutton
Andreas Neumann

From left Borys Jurgiel, Marco Hugenotbler, Carson Farmer

From left: Otto Dassau, Horst Duester, Werner Macho, Juergen Fischer

From front Horst Duester (facing away), Werner Macho (facing away), Juergen Fischer, Tim Sutton, Borys Jurgiel, Marco Hugentobler

We discussed Python and plans and possibilities for python plugins into the future.

We also discussed issues with GEOS and the possibility of replacing GEOS with a library such as CGAL into the future.

We also discussed creating a new library libqgis_analysis which would contain analytical routines that are currently written in python, with python bindings so that the existing python dialogs can be used. The advantage of this approach will be a) improved analytical performance and b) the analytical functions can be used by other 3rd party aplication.

The translation folks are going to create a german version of the QGIS home page.

Thursday 19 March, Morning

I finished getting QGIS to build under MSVC / Windows using only libs from OSGeo4W thanks to help from Jeurgen. Next step is to create a standalone package installer for Windows.

Wednesday 18 March, Afternoon:

Horst Duester gave his workshop for Introduction to QGIS and that was well received. After that, I gave my talk on the QGIS outlining the road we followed to get to version 1.0. There were a good number of folks attending the meeting. Next we had a QGIS Users meeting to brainstorm what people would like to see QGIS. The brainstorm session was attended by lots of QGIS power users which was great and many good ideas were tossed around. Here is the raw list of things people asked for in 2.0:

QGIS 2.0 Wish list:

* Better support for GEOS functionality, OGR functionality, GDAL functionality - analytical tools.
* Distribute important plugins by default:
- postgis manager
- ftools
* Plugin organisation - tagging / functional separation so you can e.g. grou postgres plugins
* Make plugins more accessible - like not having to enable python installer and merge two plugin installers

* Scale dependent rendering - use scale ranges with different symbology per range
* Support mixed geometry layers (e.g. show them as sublayers)
* Pasting rows from a postgis tables pastes binary geometry (bug)
* Deeper heirachy for layer groups
* When hovering over a group it expands the folder and doesnt collapse them
* Be able to select multiple legend layers and e.g. delete them
* WFS-T support
* WCS client support
* Undo?
* Template support in print composer (already partially implemented in 1.1)
* Classify more that one vaue into a single class in symbology and have one legend entry.
* Scale point size in graduated symbol by attribute value.
* Display PostGIS layers with not pkey even if they cant be selected or edited.
* TMS Tiled Map Server client support - when dealing with fixed resolution intervals fetch data from the next zoom level down.
* Support handheld platforms.
* For WFS features fetch only features in bounding box rather than the whole dataset.
* Support user defined Qt ui forms when editing data.
* Table joins and table manager.
* Print composer - display coordinates around edge of map
* Preserve Z and M values after editing in OGR datasources.
* Allow editing of Z and M values
* Support time series for WMS client so that you can pull down images from a certain time range.

Version 3:

* True 3D?

We will have to look carefully through the above list and see which things are doable in the time we have till QGIS 2.0 comes out. All in all it was great to have the QGIS User meeting and really gratifying to see people making good use of QGIS.

Wednesday 18 March, Morning:

I sat around at the QGIS booth for the day and chatted to various booth visitors. Silke Reimer from RapidEye (and formerly of Thuban / Intevention fame) came by and chatted bout RapidEye and the use they are making of FOSS GIS there. Several other folks came by from public administrations and the like. There is a lot of interest in migrating / adopting PostGIS as a spatial data backend and a general desire to operate mixed environment, with ESRI products connecting to PostGIS for power users and then using QGIS or web clients for the 'data browser' class of user. However there is still some question as to a) the status of GDAL connection to SDE capabilities and b) the ability to read/write PostGIS data from ArcMap without using SDE. Hopefully someone reading this will post some insight into the options that exist in the comments.

Tuesday 17 March, Afternoon:

I posted some pics of QGIS enthusiasts here: fossgis20009 - call back again there as I will add more pics as the conference continues.

This afternoon Juergen and I poked around a bit to sort out more issues towards making QGIS build under msvc with osgeo4w dependencies - including some changes I made to cmake so that geos, gdal, sqlite and postgres libs automatically get picked up from the c:\osgeo4w dir now. Still need to sort out GRASS and Python support and then I will start to look at getting the windows standalone package working again (but using msvc build).

Werner Macho and I poked around a bit trying to make his Magellan Triton 2000 GPS unit work with QGIS and / or GPS babel but it seems its a no-go since the unit uses an undocumented and unsupported protocol :-(

More later...

Tuesday 17 March, Morning:

Arrived here at FOSSGIS2009 in Hannover and already met up with fellow QGIS enthusiasts:

Otto Dassau
Werner Macho
Juergen Fischer
Horst Duester
Marco Hugentobler
Andreas Neumann

Also met with Pirimin Kalberer who showed me a WMS search plugin he developed that uses the OpenSearch API to search for available WMS servers. I suggested that he integrate the functionality into the existing WMS Plugin. Hopefully we will get a chance to hack on that a little during the week!