Re: <OK> Re: [Groff] changing .em behaviour?

[Werner LEMBERG]

> Denis makes a good point, but I wish to offer a counter argument.
>
> If the model of groff was "compatibility over all" I would agree
> with him.  But, in my opinion, groff is not aiming for compatibility
> as the first requirement.  The very fact that the line
>
>       .sp1
>
> yields
>
>       warning: `sp1' not defined (probable missing space after `sp')
>
> (unless the -C option is invoked) illustrates that groff has moved
> beyond historical troff.
>
> To me that implies that the unexpected behavior of .em should not be
> preserved.
>
> The fact the .em will act differently in compatible mode, and that
> fact will be documented, is good enough for me.

Exactly.  The compatibility mode should be as close as possible to
AT&T troff.  In `normal' mode, groff should behave as sensible as
possible.  As mentioned earlier, the bizarre behaviour is of no real
use and most likely a documented bug and not an intended feature, and
I can't think of anyone (ab)using .em that way.  Actually, it's quite
surprising that in all the years noone has ever questioned why .em is
behaving that strangely.


    Werner