this is not an IE world
The answer seems simple to me: just add to every "alternative" browser
a button to revisit the current page using the Default Megabrowser
installed in the system, and people will use small browsers most of
the time and stop complaining about no browser being able to do
interpreting wrongly-written code + etc cannot be considered, in any way,
as the "current standard for web pages", and trying to support all
that is like trying to support all formats for writing documents of
ten years ago -- straight text in many charsets, ANSI codes for
colors, *ROFF, TeX, PostScript, Interleaf, and hundreds more -- in a
single package. Some browsers try to do that, and some like IE are
almost doomed to do that very well because every commercial page is
tested on them...
It is not true that what users want is "a browser (or computer) that
does xxx and yyy"; in the end what users want is to have a computer,
and programs, that make them happy, and there are many ways to get to
that. If your way is by following all the newest hype, fine, but other
people have other ways that also work very well. Many people have
switched to Linux just because it is a joy to use, even if it doesn't
have all the features of IE, or if they will miss the dancing paper
clip. Other people will prefer tiny machines running Forth as their
OS, or videogame consoles. My way of making my friends love *nices is
to start by teaching them Emacs Lisp, and it has worked much better
than I believe than just "following the market" would. And the *nix
shells support a safe kind of plugin that you haven't considered: as
an unpriviledged user, download http://xxx/yyy.tar.gz, unpack it, read
the instructions, ./configure; make. This is much funnier than just
clicking and watching images pass by your screen.
The war is over; what you are trying to consider as war is just
illusion, you say "no" to it, and, snap, it goes away. M$'s idea of
having the same computer on every desk is not the only truth, just one
extremely biased, and pernicious, point of view.
There's a program called Kibitz (see the Expect home page, at http://expect.nist.gov/ (http://expect.nist.gov/)) that lets two people share the same text terminal in two remote machines; keystrokes from both people go to the same program, that is running in the same terminal. One of its goals was to promote remote debugging sessions.
Anyone knows if it is being widely used somewhere? I'm afraid it never got the recognition it deserves, and its development got stuck (its security concerns seem a bit old-fashioned, for example).
It is a public-domain Expect script only a few hundred lines long, and it should be very easy to change...