[GRASS5] fwd: [C.Hoegl@gmx.net: reaction to questionnaire]

[Moritz Lennert]

Finally I've found the time to analyse the questionnaire responses. Here's one that I thought should be posted here:


----- Forwarded message from Christoph Hoegl <C.Hoegl at gmx.net> -----
> 
> >  
> >  
> >  8) What are your wishes for the future of Grass (which new modules, gui, 
> >  database management, etc) ?
> >  
> The module base is pretty complete but pretty unusable for the novice.
> You have to look in the source code to get a glimpse of what specific
> modules do, and how they do it (Even some "Grass Core Developers" don't
> know of some of the "features" of the code. Part of it seems to be the lack
> of documentation of the core source (CERL-time) which was dropped more
> or less on the migration to a non-military project), which now hinders
> new coders to deploy new methods as all core routines are pretty messed up.
> 
> A pretty easy plan to make grass more feasible to the beginner would
> be to reduce the number of modules to the working core modules
> and have people download all modules they might need later on.
> 
> This would work like this.
> Splitting the tree in
> 	Import/Export	like r.in.* v.in.* r.out*
> 	Display		like d.vect*
> 	db		like db.*
> 
> And make conclusions on wether they do their core functionality and on what
> core functions they depend on.
> 
> Provide a testmodule that checks for the existance of modules and tests
> for bugs. That way a complete prerelease-check would be possible.
> Provide a list of defunct modules along with a way to circumvent the particular
> misbehaviour.
> E.g.: vector line types (boundaries, edges, lines) : Status: currently
> not yet working correctly. Use separate maps/overlays to get calculation done
> and merge only for presentations. This will save you lots of time
> and your data becomes more useful.
> 
> >  
> >  

----- End forwarded message -----