(j3.2006) READ unit ambiguity

[Van Snyder]

Bill Long wrote:
> Interestingly, there is a constraint immediately following that says
>
> C812 (R819) The do-variable shall be a variable of type integer.
>
> Is that constraint really necessary?
>   

There isn't a constraint (that I could readily find) that 
<int-variable-name> or is of type integer, so apparently C812 is 
necessary.  There is an assumed constraint (C101) that a prefix of 
<scalar-> on a syntax term requires a scalar.