Re: lynx-dev next/prev relative number command suffixes

[Jacob Poon]

On Sat, 27 Feb 1999, Laura Eaves wrote:

> > Date:       Sat, 27 Feb 1999 15:38:20 -0500
> > From: Jacob Poon <address@hidden>
> >...
> > On Sat, 27 Feb 1999, Laura Eaves wrote:
> > > > The syntax is a little awkward.  The +|- sign should be put in front of
> > > > the numbers. 
> > > 
> > > I agree, but + and - are already keystroke commands.
> > > A compromise would be to use 0+5g -- require a leading 0
> > > in order to use the +/-.  Comments welcome. 
> >
> > I think there is a slight confusion here.  You are patching F_LINK_NUM
> > command, right?  Since your current scheme describes the texts to be
> > entered in the prompt after invoking F_LINK_NUM, not the actual keystroke
> > involved for using F_LINK_NUM feature, whether the +/- keys are reserved
> > or not should not matter at all.
> 
> Yes, I patched F_LINK_NUM -- but the user needs to type
> a numeric key to reach that code in the first place.

Actually, F_LINK_NUM can be binded to any key, through the use of
lynx.cfg.  But even ignoring that, + and - keys are on the keypad, so
there is no distinct disadvantage for putting signs first.

> Would typing 0+1g be less awkward / more natural than 1+g?
> If there's enough interest I'll correct my patch to accept that syntax.

0+1g (keystroke sequence) is more consistent because:

Users will be more aware of the differences between relative and absolute
links.  If the first character is signed, then it is a relative link,
otherwise it is not.  On the other hand, by putting signs behind numbers,
users will not aware a mistype until after reading through the numbers. 
Although RPN fans will have no problem of that, many other Lynx commands
are not RPN in nature, so adding RPN syntax on F_LINK_NUM will result
reduced consistency (besides, there aren't many pure RPN calculators
compared to non-RPN ones).