Re: Documentation of transient-mark-mode is sloppy, wrong, and confused.

[Andreas Roehler]

Stephen J. Turnbull wrote:
> Eli Zaretskii writes:
> 
>  > > > email.  Sorry for the poor wording - it's late and I'm tired.
>  > > 
>  > > I would find this fairly confusing
>  > 
>  > So would I.  If you read this text carefully, it actually says
>  > something like "the region is active when Emacs marks it as active."
>  > That's hardly a definition of what "active" is.
> 
> Read a little more carefully.  It also explains how the behavior of
> Emacs changes when the mark is active.  That *is* a definition of 
> "active".  ACM did say he was tired, but he's rarely *that* tired.
> 
> How about
> 
>     In Emacs, the region is the substring of buffer text between point
>     and mark.  If the mark is not set, then there is no region.
>     (Point always exists, and the mark is normally set early in the
>     life of a buffer, and is moved, not unset, thereafter.)  Some
>     commands operate on this region, copying it, killing it, deleting
>     it, or transforming it.  [Insert list here if you like.]
> 
>     In transient mark mode, the region may be active or inactive.
>     This is implemented as an internal flag which is set implicitly by
>     commands which highlight the region (for example, "shifted motion
>     commands").  The flag's value is accessed with `region-active-p'.
>     In transient mark mode, commands that operate on the region
>     require it to be active, or they signal an error.  That is, when
>     transient mark mode is on, we say the region is @def{active} when
>     region commands are effective.  Otherwise, we say the region is
>     @def{not active}.
> 
>     When transient mark mode is off, the region is never highlighted,
>     and commands that operate on the region will error only in the
>     case that the mark is unset in that buffer, and the region does
>     not exist. 

IMHO its better to let this up to the programmer/circumstances.
It might be well useful to make user aware of region via highlighting,
preventing accidently killing and the like.

Raised was just the question if the highlighting should be called
´transient-mark-mode' - as there is no transition so far.

Why not introduce `region-highlight-mode' instead?


 We also say the region is @def{not active} whenever
>     transient mark mode is off.
> 
>     As a sort of abuse of terminology, it is often said that "the mark
>     is active" (resp. "not active"), when what is meant is "the region
>     is active" (resp. "not active").  "Inactive" may be used as a
>     synonym for "not active".
> 
> I'm probably missing or wrong on some details, but as an outline
> something like the above should do.
> 
> IMO, a large part of the problem here is confusing the mark with the
> region.  If the mark is set, you can always do `(goto-char (mark))' or
> M-x pop-mark, etc.  The mark is *not* transient.  The region is.  But
> I guess it's too late to fix that.
> 
> I swear on a stack of Bibles that the above text was written by me, my
> employer has no clue or claim on it, and I hereby contribute it to
> XEmacs (and veto its application since we don't have transient mark
> mode, thank Jamie!)  But it's now the FSF's; do with it what you will,
> my papers for XEmacs are on file.
> 
> 
>