basically I have nothing against "aunits" as long as the default
behavior remains the same .. except that it is a very small
change and we would like to have more improvements
at the same time to the Proj.4 definition language.
One of the most important is the formal syntax checking and
check against most obvious user input errors (typos etc) ..
for example normal lat / lon limits on Earth etc. If errors were
detected the separate routine would then spit out a warning
or error message. This would then warn the user in fore
hand that there is an obvious error in the input. Let say
for example having an O (letter O) instead of a 0 (number 0)
in a number.. Now the sad situation is that Proj.4 says
nothing but gives crazy results and the user is not aware
of what is the problem and the crazy definition might
end up in some library and used with strange results.
The checking could be a separated subprogram that just
does the checking and nothing more. The main program
could just omit that with such data that is already checked
once to keep it simple and fast. So the check routine would
be a separated part and would not affect any of the current
library operation and programs that use it.
It would be most convenient to have that check inside
the library since it also defines the interface language
and format. Now everybody have to write it separated
if they want to have that.. situation similar to "aunits".
The syntax checker and analyzer could be also a
some kind of a preprocessor that could also handle
the "aunits" as well .. and maybe some future additional
So if you do all that ... I have nothing against "aunits"!
> Hi everyone,
> Is there any chance you will consider my +aunits= features (ticket
> #121). I am willing to donate any time required to get this integrated.
> Proj mailing list
> [hidden email] > http://lists.maptools.org/mailman/listinfo/proj >