Re: Backspace meaning "go back" in browsers

[prev] [thread] [next] [lurker] [Date index for 2004/12/19]

From: peter (Peter da Silva)
Subject: Re: Backspace meaning "go back" in browsers
Date: 14:46 on 19 Dec 2004
> What fucktard decided that hitting "backspace" should cause your browser to
> go back a page? There are ALREADY NAVIGATION SHORTCUTS.

One of the myriad fucktards who decided that something that should be as basic
and user-configurable as keyboard mappings for graphical applications are
nestled so deeply in layers of code and macros and triggers and input chains
that not only are they not easily configurable by typical users but they're
not practically configurable by people like us.

Unless a program (application, window manager, display server, GUI, what
have you), was written in 1978 to fit in 64k or less memory, it should have
a table that looks something like this:

	Key Code	Context		Command
	--------	-------		-------
	backspace	input		erase-left
	backspace	*		page-back

And an interface that lets you change it easily and interactively. And this
needs to operate at EVERY level. Keyboard to logical key name, triggering
plug-ins, mapping keys to GUI operations, and so on. The only GUI that even
manages to do a decent job of PART of this is the oldest and most primitive
one around... the newer ones delegate it to a twisty mass of obscure hooks
like amateur macrame made from rosebushes.

I mean, for god's sake, I wrote a file browser in C on the PDP-11 that was
so crammed for space that after it started up it wrote the environment area
to a tempfile and used the 1k freed up as a file buffer. And it still
allocated 128 bytes to a table like this. There's code I wrote on Soureforge
right now that managed to cram this kind of capability into an interpreter
that had to fit into less than 40k so it was usable on Xenix in small model.

I wrote a full-screen editor whose SOURCE CODE fit in less than 10k, that
I used on modestly configured micros with 16k of RAM, that let you modify

	keytab command
	  char a => append
	  char b => back-word
	  char c => change
	  ...

	keytab insert
	  32 127 range => self-insert
	  ctrl [ => end-table
	  ...

Here we are, it's 22 years later, I'm using a computer that has not one but
two processors, from two different companies, that are each a good four years
behind the state of the art, and they can still do in real-time work that
took a dedicated high-end minicomputer ten hours a frame to complete back
when I was writing that code... it's got what may be the most sophisticated
user interface in general use on this planet... and yet there is still no
component responsible for command dispatch at ANY level.

WHAT THE FUCK ARE THESE PEOPLE SMOKING?

Generated at 02:00 on 04 Feb 2005 by mariachi 0.52