>Given an image index, plus the coshape and lower cobounds, it is tedious but
>not too difficult to calculate the cosubscript values required - the
>calculation is written down in the standard (it's the same one as the array
>element sequence).
Yes, C.10.1 - very elegant.
I can't find it in 15-007r2.
Or has it moved to the further coarrays TS?
>It is reasonable for IMAGE_INDEX to exist since the processor itself
>regularly needs to convert cosubscripts into image indices (for
>communication purposes).
Sorry, I don't understand you here.
What do you mean by "processor itself ... needs"?
Can you give an example?
> The case for the inverse function is rather
>weaker, since
>(a) the processor is not doing it itself,
>(b) the user can just write the function himself.
Ok, thanks.
Just wanted to check that somebody has thought of this.
Anton
|