Subject: Re: [coallesced replies] Re: how to create 21,780 new free software jobs (2,530 in R&D)
From: Rob Helmer <>
Date: Sat, 9 Nov 2002 18:37:49 -0800

On Sat, Nov 09, 2002 at 05:28:40PM -0800, Brian J. Fox wrote:
>    Date: Sat, 9 Nov 2002 10:45:39 -0800 (PST)
>    From: Tom Lord <>
>    Absolutely: control over the harware platform is critical.  Computers
>    need to be more like phones: cheap and swappable and relentlessly
>    standardized in this area.  Is it a rental model we want?  A shelf of
>    endorsed hw at CompUSA?  a page at Dell's web site?
> Of course, this model was in fact the rage of 1999.  Network
> "appliances" which were cheap or free, but were service-locked to an
> ISP, and could have their software updated at any time by that service
> provider.
> It failed, mostly because people didn't want to be service-locked.

I think this is a very valid criticism.

I've always liked the idea of the network appliance ( I bought
one of those I-Openers for my grandfather, they ended up going
out of business, but Earthlink picked up the service end so
it wasn't so bad. He still uses it, it works extremely well 
for his purposes ).

I never saw one that I was really interested in though. There
was no standardization, very few programs available, etc.
The I-Opener was nice in some ways, but completely closed.

People started figuring out how to boot Linux and NetBSD
on it, so the I-Opener guys started gluing the chips
and the screws in so you had to be really careful to
not break the thing if you wanted to modify it.

You definitely could not use them as an ISP with anything
but their software.

I think the whole reason this is applicable as an FSB but
not neccessarily as a proprietary business is :

* the "service lock-in" aspect is moot, anyone can do this.

* there are many decent free desktop applications. there are
  tons of alpha/beta quality packages that could be developed
  further, and there is great potential for proprietary 3rd party
  apps to have a standard way to integrate into any "Linux" system.

* there is already standardization movement in the "Linux Desktop"
  area, see KDE/GNOME's usability projects, Debian desktop project, 
  LSB,, etc.

This last point is important because it is not someone like
Mandrake, Suse or RedHat doing this work on their distro and
selling it as a value-add - it's an attempt to standardize the
basic desktop peices that they all use, rather than the 
haphazard stuff we've seen in the past.

All of the network appliance vendors needed some kind of
lock-in, either on the network service side or the OS
or the applications. They weren't interoperable with
eachother. Many of them had abysmal HTML standards support
in their browsers, even ( probably not on purpose ). 

Lock-in was part of their business plans. There wasn't
an attempt at a free, standard platform for network appliances,
that they would divvy up the users the best they could,
with some being more knowledgable and hence better
at providing support for each other ( not that I'm
aware of anyway ).

It was winner-take-all, and they ended up all being

The biggest problem is that modern-day computers need 
system administration, they need to be patched, and these
activities get in the user's way. O/S providers need 
money to do this at a fundamental level, so they charge
for new CDs.

Perhaps if everyone took their computer in for a "tune-up"
like they do ( or should do ) to their car, it wouldn't
be such an issue, but this does not happen, for one reason
or another.

Desktops are also notoriously unreliable, what with crashes,
data loss, mysterious errors that "go away" on their own,
viruses, etc.  many users are afraid to really use their computers, 
and the problems only get fixed if they maintain their computer,
make and test backups, and use good security policies.

I think all of these could be addressed in a way that's
transparent to the user, they just *use* the computer.

Not to say that these things won't happen, I'm sure
they will, to some extent. The differnces is that
they can be dealt and fixed for the users without
them having to get involved.