[GRASSLIST:2954] switching to GRASS

[Aurelien]

Hi,

I have yet only lurked into GRASS briefly, but now I might consider it 
with more attention. I'd like to ask you about a possible migration path 
from a custom-developed java/mysql solution to GRASS.

Until now, I have imported data from a proprietary format (TAIF) into a 
mysql database. The aim is to generate maps for the web fast, using 
Flash (swf) as an export format. The current solution does that very 
well, but I'd like to open up for more possibilities and that's why I'm 
turning to GRASS.

Specifically, a benefit of the current layout is that the rendering 
process is very fast, and doesn't require any application support: I 
simply fetch the optimized data from the database and use my own flash 
generation library to send the map to the client.

A VERY IMPORTANT point is that I need to optimize the edge/face data for 
the vector format, as the number of points to draw is in direct 
proportion with the amount of data to send through the wire.

The first thing I do is import the geometric object. For example, I 
merge points that belong to edges into an encapsulated "edge" object.  A 
lot of crap must be done on the road edges, that originally come with a 
very fine-grained specification scheme. The roads are thus merged 
according to what information I really need.

Are such things already implemented in GRASS ?

Then I wrote a quite complicated road-labelling algorithm. Again, I 
wonder if such things already exist in GRASS, and if not, I might 
contribute something also. For roads that couldn't be automatically 
labelled, I wrote a simple GUI that allows for manual repositioning. 
That's something I'd rather not want to rewrite, so at least the db 
backend should keep some of its present structure.

Any input greatly appreciated !

Aurélien