[GRASS5] v.in.ascii

[Markus Neteler]

On Sun, Jul 03, 2005 at 07:54:19PM +1200, Hamish wrote:
> > Do we need topology for point data? If yes, what is in the topo file
> > for points? Would it make sense to implement -b option (no topology
> > built) for v.in.ascii as Radim has suggested and then change the check
> > for level 2 in v.surf.rst main.c from fatal error to warning?
> ..
> > Or would it be better to look at V_build and modify it for point data
> > so that the topology for points (which I must admit do not know what
> > means) does not use excessive memory? We need spatial index for point
> > data but topology is not quite clear,
> 
> 
> I prefer good defaults to multiple implementations.
> 
> i.e. either build topology for points or don't. "-b" and two levels of
> point data maps is confusing.

I agree. What about "if only points in the map, then no topology"?

> Alternatively, no -b but if the number of
> points found during the scanning step of v.in.ascii is greater than say
> 100000, issue a warning that "topology will not be built, use v.build if
> you want, etc." and skip the build step. But I expect that message will 
> be lost in all the chatter that the module produces and it is non-
> intuitive that this might occur..?

This is probably confusing as well. The system should decide, not
the user. 

Markus