I'm definitely agreeing here, but thought I should point out this can be
overstated too (the similarity) or taken the wrong way if your a typical
end-user looking for a laptop to use with free software.
While the chipsets do matter a lot there are other issues with determining if
a laptop is going to work well with free software operating systems. I'd like
to point this out as its customary to think that it's just a matter of
getting the right chipsets.
One example (not laptop specific) is the Datel Wireless 'N' Networking
Adapter For Xbox 360. This is a USB N adapter that uses the same
AR7010+AR9280 chipset that we're using. However its not compatible with free
software operating systems and probably will never be. It's not just a simple
ID issue which can be corrected with a bug report and patch. There were some
serious enough changes to the reference design that it doesn't work and won't
work as the changes were not made public.
Digital restrictions. Dell, HP, IBM/Lenovo, Toshiba, Apple, and I believe
Sony use digital restrictions to prevent users from replacing the wifi chip.
Apple's newer systems apparently ship with proprietary wifi cards rather than
use digital restrictions to prevent you from replacing the wifi chip
There are many other potential issues on systems from companies who aren't
using digital restrictions (as far as I'm aware anyway). For example there
are issues on at least some (many actually) laptops where the on/off switch
for the wifi does not work. The design is dependent on a driver which
activates it and by default is in an off state. While you can sometimes get
these systems to work it requires a proprietary operating system, the
installation of the MS Windows driver, turning the wifi on in MS Windows, and
then rebooting. Obviously not an ideal solution for free software users.
There are also power management issues. This can sometimes be related to poor
BIOS implementations and non-compliance with standards. There are some
work-arounds on many systems, but it can be particularly difficult to figure
out for people who are not already familiar with working around these types
of issues.
Other issues I like to bring up are IDs not being in the driver even if the
chipset is otherwise supported. Basically what this means is the driver
doesn't know to load for the card even though it would otherwise support it.
These issues frequently find themselves resolved down the road, but if your
buying enough random hardware is a problem you'll run into occasionally.
There is also another issue I like to talk about. Some types of devices have
multiple chipsets (laptop boards are a good example, but this isn't only
laptop boards as my example will illustrate). There is a bluetooth wifi combo
card in many laptops. The chipsets in one of these cards is a combination of
different atheros chipsets. One is for the wifi and one is for the bluetooth.
The wifi part is completely free and no propritary firmware is needed. The
bluetooth on the other hand is not free as it is dependent on non-free
firmware. The confusion comes in because many people are using distributions
with proprietary software alongside 100% free distributions like Trisquel.
When a distribution which includes proprietary firmware is booted the
proprietary firmware is loaded for the bluetooth chipset on the card. When
the user reboots the firmware is still loaded and as a result it appears to
work in Trisquel. Simply working in Trisquel is not really enough to conclude
it's free software friendly as a result. Even if you power off the computer
the firmware may still be retained in memory on the card. It may take a bit
of time for it to dissipate and return to a non-working state in Trisquel.
Anyway- a lot of people have made the claim these cards are free software
when they are not.
In any event there is definitely value in the proper testing, support, and
documentation aspect which results from a commercial operation. Something
that would be near impossible to achieve in an h-node style hardware
database. Prior to h-node there had already been dozens of such databases and
my involvement with a company for which I will not name (internship at a
company that was doing it *all wrong*) in this aspect of QA and support is
what lead me to realize we (ie the larger community/commercial
distributions/etc) were doing it wrong. Both in terms of not realizing the
harms from supporting non-free software and the way in which hardware support
needed to be done (ie to make it easy enough for the masses). Databases just
don't work. Respect Your Freedom certification does because it ensures the
advertising is such that users can identify the product supports free
software and has not been changed in such a way that it no longer does. If
the product changes in such a way the model numbers and similar information
much change- or the certification will be revoked. It essence it ensures
users that the hardware they are buying is in fact the hardware they are
intending to buy and has not been substituted with another product (this is a
major major issue with hardware databases).
