> The question is whether we should enforce it or not. Enforcing it makes
> it consistent, but harder to implement generically, especially for
> booleans.
>
> I will ask this question in engineering meeting later today.
OK. What we decided in the meeting is that
- the variable substitution case is atypical and can be handled by the generic
set command.
- command implementations can/should use typed parameters as/when required.
(This is contrary to what this thread started with).
I will make a few changes to commands to comply with this.
Thanks,
Kedar