[Qgis-developer] GSoC ideas

[Alexander Bruy]

Hi,

2012/2/27 Martin Dobias <wonder.sk at gmail.com>:
> fTools should probably experience some core/gui decoupling first - I
> remember the individual tools contained a mixture of algorithm logic
> and gui stuff handling (progress bar etc). After that, the tools
> should ideally implement a common interface, to allow easy running of
> the tools from API (instead of having to use slightly different calls
> for each tool).
As I can see some algorithms already ported to C++, see QgsGeometryAnalyzer
class. But currently this implementation don't uses threads and memory
layers.

> For that we need a simple processing framework -
> during last year's GSoC such framework has been developed as a side
> effect - but it still lacks various important things such as: show job
> progress, cancel running job or asynchronous job execution. Therefore
> a summer project could involve adding missing functionality, porting
> from python to c++ for inclusion in qgis_analysis library and possibly
> adding few modules from fTools to demonstrate the capabilities.
Hmm... If I remember correctly, when Processing Framework project
was started we also have discussion which language to use Python or
C++. And decision was to use Python...

Anyway, I agree that porting fTools to analysis lib with or without processing
framework will be great.

-- 
Alexander Bruy