[GRASS-dev] [INFO] Line simplification algorithm

[tlaronde at polynum.com]

FWIW, I seem to remember a thread on your list about the line
simplification algorithm, and the need of implementation of 
Douglas/Peuker kind of algorithm.

ISTR too that when I used GRASS GPL (back in 2003), v.prune(1) was doing
its job the worst way possible (making local simplification, with a
sliding pair, i.e. cumulating errors by comparing a vertex against the
previous one and discarding the previous if it was in the threshold, and
then comparing with next, leading to the disappearance of the curves).

Looking at what I have in KerGIS (that is the original program with Dave
Gerdes implementation) I'm puzzled... since this _is_ a Douglas/Peuker
kind of algorithm with some slight modifications that make sense.

So if you want it, just go back to the origin (adapting to your new
vector engine).

One question remains: why and who discard a working program to replace
it with a disaster? 

To be honest, working with the original sources, I know who I can trust
(developer or programmer name)---and I know I have (had) only
cosmetic work to fix the program---and who I can highly distrust---and
who has made such a poor job that is eating my time not in proportion
to his contribution (I will have spent 90% of the time fixing 10%
of the sources and know I will have to redo later everything from
scratch). Blunders may occur---been there, done that. But there
are very suspicious blunders...

YMMV,
-- 
Thierry Laronde (Alceste) <tlaronde +AT+ polynum +dot+ com>
                 http://www.kergis.com/
Key fingerprint = 0FF7 E906 FBAF FE95 FD89  250D 52B1 AE95 6006 F40C