Re: lynx-dev Re: ^Ve considered harmful

[Klaus Weide]

On Mon, 15 Feb 1999, Kim DeVaughn wrote:

> On Mon, Feb 15, 1999, Philip Webb (address@hidden) said:
> |
> | my vote is to retain  ^v , which is consistent meaning
> | `treat the following character as a command, not as input',
>[...] 
> 
> | but change  e , which (as KW points out) prevents editing the doc's HTML:
> | a quick look at the Key Page shows  a d E r s z  all lack meanings
> | which could ever be wanted in this situation (in a textarea):
> | pick one to mean `rouse my editor to add to this textarea'.
> 
> I think you're misunderstanding something.  The char following the quote/
> escape char (^V) must be a key mapped to a lynxactioncode,  In the above
> list you mentioned:
> 
>  a == ADD_BOOKMARK
>  d == DOWNLOAD
>  E == ELGOTO
>  r == DEL_BOOKMARK
>  s == INDEX_SEARCH
>  z == INTERRUPT
> 
> by default.

Well, those could be made to have a different meaning if invoked while
on a text field, just as e = EDIT is now.  At least 'E' and 'r', even 'z'
have the advantage that their normal meaning is never meaningful while
on a text field.  Of course they are less mnemonic, and I don't like
either doubling the meaning of a completely unrelated command or using
^V at all... 

> I like the idea of finding a special key that would invoke the editor in
> a *single* keystroke, and in fact that's what I wanted to do in the first
> place.

That would be ideal.

> Unfortunately, *finding* an "acceptable" binding is not an easy thing to
> do, given widely varying personal preferences (no, I WILL NOT give up ^E
> moving to the EOL, etc :-) ), but I *will* review the current sets of
> LYE_bindings again.

If we only find one that is "acceptable" for "nearly everyone",
^V-EDITTEXTAREA or ^V-DWIMEDIT could still be kept as a fallback.
 
> The real solution, IMO, is probably to allow the user to remap the hard-
> coded LYE_bindings externally, just as they can now do with LYK_bindings
> via the KEYMAP lynx.cfg statement.  *That* is something I'd like to see
> irrespective of the current discussion about ^V, editors, and such.
>
> I dunno if the recent work by DK and KW will aid in doing something like
> that or not, 

It was mostly fixes, really

> as I haven't had time to look at their work, but since they
> have been looking into the keybinding code in-depth anyway ... :-) ...
> 
> More later ...
> 
> /kim
> 
> PS:  WRT the Subject: of this thread ... Dijkstra was wrong.

Was anyone right?

   Klaus