[Battlemesh] TP-Link FCC deal
daniel at makrotopia.org
Tue Aug 2 13:17:46 CEST 2016
On Tue, Aug 02, 2016 at 12:18:20PM +0200, Elektra wrote:
> Hi all!
> Very well said, Daniel. I'm new looking for the "like" and "retweet" button in my email client.
Thanks. I reckon the "Forward" button of an email-client is as
close as it gets to "like" or "retweet" in this case :)
> One point: I don't tend to assume conspiracy if stupidity is the most likely motivation ;)
I'm also not assuming conspiracy to be the cause; opportunism, greed
and short-sightedness explains it all very well. The ruthless attempt
to even look good while trying to make a buck out of that disaster is
what makes it look like conspiracy, I guess.
> Am 2. August 2016 11:49:03 MESZ, schrieb Daniel Golle <daniel at makrotopia.org>:
> >On Tue, Aug 02, 2016 at 07:14:15AM +0300, Jonathan Morton wrote:
> >> > On 2 Aug, 2016, at 01:05, David Lang <david at lang.hm> wrote:
> >> >
> >> > If the post later in the thread is correct, the fact that they got
> >in trouble for their stock firmware allowing too much power when they
> >were trying the claim that they were preventing that by blocking open
> >firmware is very interesting to see.
> >> It’s even possible that this practical demonstration that closed
> >firmware was no more trustworthy than open was a final deciding factor
> >for the FCC.
> >... because people were able to change the country code and thus use
> >channel 12 and 13 which aren't part of the FCC's ISM band.
> >If they expect this type of deliberate transgression to be prevented
> >by $vendor while still allowing the user to use 3rd-party firmware,
> >that would really force device-makers to use DRM-techniques (like an
> >efuse to only allow setting the country code *once* at the time of
> >import...?) *and* isolate the whole wifi subsystem either in hardware
> >(Marvell, Intel, ...) or by introducing a $vendor-controlled hypervisor
> >(like imgtec/MIPS suggested). Both is very, very ugly from a FOSS point
> >of view, devices which were running (almost) 100% free software (and
> >thus allowing for indepenent review) will end-up to be as 'free' as
> >in Android or other Open Source jails surrounded by proprietary walls.
> >I'm disgusted by how this debate is being used by certain players to
> >get rid of competitors, pretending this could be solved by being 'well
> >prepared' or by making 'better products' or any other random claim a
> >marketing fart could smell like. Ok, I'm less disgusted by marketing
> >departments of $vendor to try that, because that's their job and more
> >disgusted by tech media outlets to buy that (I hope they were at least
> >well paid for being part of that campain).
> >I reckon this whole mess could have been prevented if vendors would
> >have populated the calibration partition as intended by the chip
> >vendors instead of ignoring it in order to not produce for a specific
> >target market and have the user set the regulatory domain herself.
> >In the end, the whole story is quite detached from reality, because
> >deliberate transgressions will always be possible by modifying the
> >hardware (think: crazy antennas) and cannot be prevented in software
> >It's like asking car-makers to build cars which automatically respect
> >the speed-limits -- and oops, in order to do that they'll have to
> >track and transmit everybody's location, make it impossible to access
> >the engine and other critical parts and give an excuse for a lot of
> >uglyness to the dominating players. Half-dying dinosaurs like Ford and
> >GM could bash on Hyundai for not being prepared for the legal
> >requirements early enough and lobby to have their products banned.
> >Actually, the car-industry recently provided a very good exmaple of
> >what happends if you trust legal compliance to $vendor and allow them
> >to obfuscate the code...
> >Battlemesh mailing list
> >Battlemesh at ml.ninux.org
> Diese Nachricht wurde von meinem Android-Mobiltelefon mit K-9 Mail gesendet.
> Battlemesh mailing list
> Battlemesh at ml.ninux.org
More information about the Battlemesh