(j3.2006) components that are data objects?

[Malcolm Cohen]

Michael Ingrassia said:
> NOTE 15.12 [398] says
> 
> The syntax rules and their constraints require that a derived type
> that is interoperable have components that are all data objects
> that are interoperable.
> 
> 
> Is there something askew with our language here?

Yes.  It's only a note though, it's not normative.

>   I thought that
> data objects were constants, variables, or subobjects of constants.

Yes.

> So how can a component be a data object ?

When it's a structure component.

>  You might even have
> a derived type for which no variables are ever declared!
> 
> What's the proper usage?

"objects" -> "entities"

> Is "derived type ... have components"   is wrong?
> 	Types don't have components, structures do.

Derived types do so have components!  See the first sentence
of 2.4.1.2 "Derived type".

> Is "components that are ... data objects" wrong ?
> 	Components aren't data objects.

Yes, it's a one-word typo.  I'll fix it in 07-007r2.

Cheers,
-- 
........................Malcolm Cohen (malcolm at nag-j.co.jp), Nihon NAG, Tokyo.