Re: Unified name properties

[Keisuke Nishida]

Dirk Herrmann <address@hidden> writes:

> Determining a primary name in all these cases is difficult, I think.  And,
> I don't want people to change their coding style just to comply with how
> primary names are derived from the code.  If something like that would
> happen, it would a very bad thing IMO.

I agree.  I think we don't need to guarantee what object-name really
returns; it just returns some name that can be used for debugging
purpose.  That will be enough in many cases.  Instead, the user
shouldn't rely on the name returned.  If the user needs some specific
name associated to a particular object, she should use another procedure
or write her own interface.  Probably we should name object-name
differently, like %object-name.

-- Kei