[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: GPL version 3

Hash: SHA1

On Thu, Apr 30, 2009 at 09:44:23AM +1000, David Cottrill wrote:


> The condition of allowing users to install a different operating system
> seems a lot less stringent than allowing users to access the underlying
> system.
> The underlying system is affected by GPL-3 and so users can not be
> hindered in their access to it.

Yes, that seems to be one of the purposes of the GPL-3.

>                                 This is the crux of my problem - the
> underlying system is not user friendly, nor can it be made so to satisfy
> usability tests (can users use it without being able to break the
> system).

That looks like a straw man to me. Resorting to the well-loved Slashdot
car analogy, I'm allowed to buy a new car and a set of wrenches and go
wild on it. If it doesn't work afterwards (most probably it won't, given
my limited abilities) -- is the car to be declared non-user-friendly?

Of course not!

There must be some clearly visible line which conveys to the user: "here
be dragons. You are on your own from here on". If you are extra
friendly, you may provide some kind of undeletable firmware which knows
how to re-load a pristine state. If the user loses configuration or
tinkering -- tough luck: she trespassed the "dragons" line, remember?

As a device provider you get something in exchange: a vibrant community
of tinkerers which may be a source of endless cool ideas for your
product (how many have bought a Linksys router just *because* it's so
hackable? I know I have; on the other extreme: watch those iphone folks
doing cool things _in spite of_ Apple trying to prevent that and
extrapolate the power this community would have if unhampered by Apple).

> I could put in a generic option to permanently disable the closed source
> parts of the box if users insisted on using the underlying system

I don't understand this part. Why should you "disable the closed source

This is GPL-3, so the "closed source parts" are well-separated
(aggregation, no linking), e.g a user-space program (say Flash, to put
an example) would be OK, a kernel module is alread gray area (I wouldn't
risk that). So... why disable?

If the user changes the environment that much that the closed-source
parts won't work -- tough luck (remember the dragons?).

>                                                                   but
> that to me is a poor option. If I didn't permanently disable the closed
> source parts then we would be forced to support devices that had unknown
> modifications - a worse option.

Just re-flash (see above) -- of course with the user's consent.

To me it boils down to "can we draw a clear line such that the user
really knows when she is leaving the 'within support' area"?

Of course, you get extra points if you are able to encourage a community
to explore the area beyond!

Here [1] is one of my favourite quotes in this context. Enjoy.

[1] <http://article.gmane.org/gmane.comp.handhelds.openmoko.community/17428>

- -- tomás
Version: GnuPG v1.4.6 (GNU/Linux)


Reply to: