[GRASS5] The status of 5.0

[Roger Miller]

On Thursday 21 March 2002 18:54, Eric G. Miller wrote:

> The v.subj thing is a hack around for poor attribute support.  I thought
> it had been disabled, but apparently not.  I don't think it's worthwhile
> to put much effort into it.

If you scratch v.make.subj then you also need to rewrite v.merge.  I suppose 
you could also scrap v.merge, or maybe someone could just right up the 
documentation for how a user could right a "subj" file required by v.merge 
without actually using v.make.subj.

> > > 6. v.proj was useful only because I was able to use the version that I
> > > rewrote to perform the nad27->nad83 conversion.
> >
> > Roger, maybe you can share the code with us? Datum transform is a
> > highly wanted issue...
>
> Markus, other than grid based transformations, there's already the code in
> the coordconv library which will do datum transformations via a couple
> different methods.  Problem is, it's not hooked into the general
> projection code in gislib, there's something of a mutual dependency
> issue with the two.

The conversions between nad27 and nad83 are necessarily "grid based," if I 
know what you mean.  There is no datum shift formula because it isn't just a 
datum shift.  As a result, having the nad27<->nad83 conversion (while 
essential for those of us in North America) does nothing to help any other 
datum conversion problem.

Some months back I modified the code for r.proj, v.proj and s.proj so they 
would automatically and (almost) silently complete the conversion whenever 
the location data requred it.  I offered it on the list at the time and there 
was precious little interest, so I didn't find time to clean up the lose ends 
and send it in.

The changes aren't just a change in *.proj modules.  There is also a change 
in the grass library function that uses the proj library.  Also the 
conversion tables need to be stored in the GRASS directory hierarchy 
somewhere.  I put them in a subdirectory under $GRASSROOT/etc.  And more, I 
didn't use the most recent version of the proj library; more recent versions 
use data tables that are valid farther north in Canada then are the standard 
conus tables.

Just the same, the programs do work and they are available.  However,  there 
should probably be some discussion of where to put it and someone else needs 
to put it there, because I don't have CVS write access.


Roger Miller