There is no good answer, except that the existence of the d.
commands makes the -i option inconsistent. For example should a
-i option be added to g.region to do d.zoom (and get rid of
d.zoom). Ideally r.what and v.what and the like should accept
keyboard event and mouse event transparently (with out having to
specify a -i option), but this impossbile with the basic design
of GRASS.
I place my vote with d.what.vect and v.what (no -i). Cloning of
code is, while not good design, necessary since the actual
"extraction" of information from the vector file (or raster in
the case of r.what) is embeded in the module itself rather than
being placed in a library routine which both v.what and
d.what.vect could use.
|In this second version, I have removed all mouse support to make it
|consistent (in this regard) to r.what. Thus, it is no longer a
|replacement for d.what.vect. This was done with ifdef's, so it is
|easy to switch it back. While it's good to maintain consistency, I
|was trying to keep repetitive code and the number of grass commands to
|a minimum.
|
|If someone has more time to spend on this than myself, I would
|recommend that
| 1. v.what be modified to include support for multiple vector files AND
| for the mouse (the latter is just ifdef'ed out right now),
| 2. mouse support be added for r.what
| 3. d.what.vect and r.what.vect be removed.
|Of course, these are just my opinions. This would be a good way
|for an aspiring programmer to get his/her feet wet [hint, hint],
|because the program design and most coding is already done.
|
|Then again, this may go against keeping display commands prefixed
|d-dot, so maybe it's not such a good idea... anyone have any thoughts
|on this?
|
|--Darrell
|
|
Michael