Weekly Engineering News 37/2008
werner at openmoko.org
Thu Sep 18 19:34:27 CEST 2008
Mike Montour wrote:
> For what it's worth, I would rank your #3 ahead of #2 by a large
> margin. It is very important for users to be able to upgrade their
I fully agree. That policy never made sense to me.
One big issue of is that this goes in the opposite direction of what
hardware manufacturers are doing: old school hardware design kept
complexity low and involved a lot of testing, because any major flaw
meant that products using that components were broken.
Upgradeable firmware changes this dramatically, since products can
be fixed in the field. Even many hardware problems can be solved or
at least worked around in software (e.g., think of our SD vs. GPS
issue). This means that components can ship as soon as the basic
hardware has been verified, but all the rest of the functionality
may take a while to mature.
And, of course, there is a natural tendency towards making things
more complex as well, which increases the probability that a
firmware upgrade will be needed. Furthermore, in communications,
security against attacks carried out over the network becomes an
issue, particularly if these attacks affect a large class of
devices and thus make them an attractive target.
If Openmoko deny themselves and/or its customers the possibility
to perform firmware upgrades, but hardware vendors expect such
upgrades to be needed in the lifetime of a product, this has the
- we cannot use new chips, because their firmware hasn't stabilized
- once firmware has stabilized, the focus of its maker has likely
shifted to newer products, giving us less support
- the time in the chip's life cycle between the hardware being sound
and the firmware being stable cannot be overlaid with our own
development process, thus shortening the effective lifespan of
- we have to prepare for recalls or major "send in" repairs
- customers will have to consider the risk that Openmoko will not
be able or willing to perform a firmware upgrade the customers
So I'd be very happy to see that "no firmware changes" policy go.
I actually wonder if there wasn't some miscommunication at its origin.
After all, many a hard disk's firmware and even the microcode of many
CPUs can be upgraded, and I haven't heard the FSF call for a ban of
processing or storage ;-)
More information about the devel