Re: info-mode needs history navigation.

[Eli Zaretskii]

On 3 Nov 2001, Kevin A. Burton wrote:

> 1. remove Info-last and create two new functions and bindings including
>    Info-history-forwards and Info-history-backwards
> 
>     - PROS: elegant and straight forward.
> 
>     - CONS: breaks backwards compatibility and requires users to learn new key
>       bindings.
> 
> 2. Keep Info-last and write a new function which handles forward navigation.  
> I
>    don't know what to call this new function (any suggestions?).
> 
>     - PROS: keeps backwards compatibility, preserves key bindings.
> 
>     - CONS: leaves us in an ackward position because the alternative nav
>       function (Info-first) is little confusing.
> 
> 3. Keep Info-last but this will be an alias for Info-history-backwards and
>    pursue case 1.
> 
>     - PROS: elegant and straight forward.  Keeps backwards compatibility,
>       preserves key bindings.
> 
>     - CONS: ???
> 
> I think this pretty much sums it up... I would prefer ot implement use case
> 3...

I'm afraid I don't really see the big difference. ;-)

The real issue is this: Info currently maintains the visited places in
stack-like fashion.  So if you press `l', the last place is popped off
the stack, and is thereafter gone.  So you cannot implement
Info-history-forward; you need to change the way visited places are
stored.

When this is done, it doesn't make any sense to keep the old
stack-based approach and the new one.  We simply rebind `l' to the new
command, and be done with that.  Voila--users don't need to learn
anything new, if they don't want to: `l' works as before; while those
who do want a new feature will use the new Info-history-forward
command.

Am I missing something?