Re: lynx-dev "sticky" things, and a half-serious proposal

[Klaus Weide]

On Sun, 17 Oct 1999 address@hidden wrote:

> On Sun, 17 Oct 1999, Klaus Weide wrote:
> >> As much 
> >> as I try not to use it as an example, this is how it works in the GUI 
> >> world -
> >> left arrow in an edit field does nothing, but you can still tab to other 
> >> fields to type in.  My "modal but not noticable" discussion above is really
> >> just emulating what I consider reasonable user-focused behavior (do what I 
> >> mean
> >> not what I say).  lvirden has already said he doesn't think default 
> >> behavior
> >> should be changed.
> 
> Ok, I realize now I've lost this whole argument.. I still hope that this 
> capability can be configured in (or turned off without recompile) -- that
> is, not ask me a question nor accidentally go back a page.

It isn't a technical problem to provide yet more behaviors, it is more
a question of configuration syntax complexity.  (And, probably, "Is
just one key worth all this special attention".)

Well, I have come up with the ultimate option:

# The LAKABOFTIF option controls controls how lynx behaves when a
# Left Arrow Key is pressed while the text cursor is at the Beginning
# Of a Form Text Input Field (in line-edition mode).  The normal behavior
# of the Left Arrow key is to return to the previous document in browsing
# history (PREV_DOC). This option can be used for setting various different
# behaviors in the specific situation, to prevent accidentally leaving
# the current document when you didn't mean to.
# LAKABOFTIF takes directives in the form 'action=condition', several of
# these can be given separated by comma.  Condition is one of:
#   NEVER            - never do this (same as omitting 'action' altogether)
#   ALWAYS           - always do this
#   IF_CHANGED       - if the current text input field was edited
#   IF_NOT_EMPTY     - if the current text input field is empty
#   IF_FORMS_CHANGED - if *any* form fields in the current page were changed
# Action is one of:
#   PROMPT - Lynx will ask what to do, giving a choice of going to the
#            previous document or staying in the input field.
#   IGNORE - The key is ignored, and the cursor remains in the input field.
#   BEEP   - The key is ignored as for IGNORE, additionally lynx tries to
#            produce and audible warning.
#   COMMIT - The last editing changes (if any) will be committed, instead
#            of being forgotten, even if lynx exits from line editing to
#            return to the previous document.  This is useful in case you
#            return later to the current document while it is still in
#            memory, for example by following a forward link again.
# If several 'action=condition' pairs are given, [ ...to be filled in... ]
# If no LAKABOFTIF option is specified, the default is to PROMPT
# if the input field has been changed, and to never IGNORE or BEEP.
# If the option is present but empty, i.e. just 'LAKABOFTIF:', The Left
# Arrow key always has its PREV_DOC meaning (or other, if changed with
# a KEYMAP options).
#
#LAKABOFTIF:PROMPT=IF_CHANGED

Do you like it? :)
Speak it loud several times: lakaboftif, and you'll never forget it for
the rest of your life...  (seriously, I find it easier to remember such
an artificial word than which-is-which among STICKY_INPUTS and
STICKY_FIELDS.)

> >But lynx is really quite different form the GUI world.  With mouseclick-
> >browsers one doesn't need the Left Arrow key for navigation between
> >documents, but in lynx it is *the* way for going to the previous document.
> 
> But command-arrows _do_ go back pages, and that's the equivalent of what
> you're doing in Lynx with arrows.

If you *want to* use something like "command-arrows" (not sure what that is,
probably because I am not a Mac user...) for navigation, then it should be
possible to tweak lynx to never use noncommand-arrows for navigation.
It may already be possible, with the right set of KEYMAP: and ~/.lynx-keymaps
settings.  That would depend on how lynx receives the "command-arrow" keys.

I would just find it very annoying to always have to press some modifier
key with the cursor keys, but that's just me (and all other "traditional"
lynx users...).

> Yes, I don't want people stuck in a page -- isn't this completely broken
> HTML though? 

It would be inconvenient (and inconsiderate), but not necessarily broken
in a technical sense.


   Klaus