[GRASS5] Future plans for GRASS code (was: How to get monitor background color?)

[Carl Worth]

Glynn Clements writes:
 > Certainly, Carl is right that functions which accept pointers to
 > "output" arrays (ones into which results are written) should also
 > accept a size argument. But then how to deal with overflows? 

[...]

 > Better to use G_store() and return a pointer; that way, the worst
 > case is a memory leak.

I agree this is probably the right approach to take.

 > Also, I suspect that this is far from the only place in GRASS where
 > this occurs. GRASS is littered with fixed-size buffers; in most cases,
 > it's just assumed that the buffer will be large enough.

Yes, this does appear common. And fixing it means changing the API and
changing *lots* of code, but it could prevent some errors that would
be very difficult to track down. Another nice benefit would be that it
could eliminate some arbitrary length limits that currently exist in
GRASS. Limitations like this can be particularly annoying, (eg. I
really wish that I could use much longer names for my locations).

As long as we're talking about incompatible changes to the code
base... What other big changes are planned for the GRASS code? Here
are some I would like to see, (I know that I have seen a few of these
mentioned on the list before).

	Elimination of arbitrary size limits
	Revamped build system
	Elimination of "single-session" nature of GRASS
	Better support for temporal data
	Better GUI, (particularly for a handheld with a small display,
		     a 1-button pointer and no keyboard)
	Others?

Are any of these changes in progress or at least planned out? Are
there web pages outlining things like this?

I'm currently "under-the-gun" to get a working system up using GRASS
for a big demo at the end of November. After that is over, I should be
in a position to devote a lot of time to GRASS development. I have a
particular interest in the temporal support and the improved UI for
handhelds, but I can imagine myself responding to an itch strong
enough to start tackling the others -- unless someone beats me to it
of course :-).

-Carl

-- 
Carl Worth                                        
USC Information Sciences Institute                 cworth at east.isi.edu
3811 N. Fairfax Dr. #200, Arlington VA 22203		  703-812-3725