Concerns about Windows Vista implications for Linux/Ubuntu

[Constantine Evans]

Karoliina Salminen wrote:
> Hi,
> 
> You may have read this already and it might be old news, but what do
> you think? This is being discussed on EFFI's (Electronic Frontier
> Foundation Finland) list currently. I haven't double checked &
> verified any of the references of the article and can not know for
> sure if it is 100% true, but might worth be checking out if you
> haven't done it already:
> 
> http://www.cs.auckland.ac.nz/~pgut001/pubs/vista_cost.txt
> 
> Particularly concerning detail can be read below:
> "Elimination of Open-source Hardware Support"
> 
> Also "Denial-of-Service via Driver Revocation"
> 
> In other words, the article claims that first of all, to meet the
> Vista requirements, hardware vendors must not release open source
> drivers and they must keep anything else about specs confidential but
> the marketing specifications listed on their web sites for consumers.
> And secondly in the case they do release the info of some particular
> card or chip type, Microsoft can at any time disable the card (when
> next automatical update occurs) which has Linux support or which is
> being used in Linux with reverse engineered drivers done for Linux
> which essentially removes the card from market which could mean (if
> the above analysis is true at all) hard time (even harder than now) to
> get any support for any Windows compatible hardware for Linux,
> especially if the hardware vendors get scared and stop releasing even
> binary drivers for Linux not to speak of yet alone the open source
> drivers. I hope the reality doesn't turn out to be this bad.
> 
> Anyhow the question is, how is Linux and particularly Ubuntu (which
> somewhat prefers open source drivers) going to deal with the driver
> issue if the vendors will increasingly be closed and further away from
> openness and Linux thanks to the obviously deliberate "Windows-only"
> locking implemented by Microsoft?
> 
> Best Regards,
> Karoliina
> ---
> http://www.karoliinasalminen.com/blog

I had written a rather long response in which I had asserted that you
were misinterpreting the meaning of that section, which I believed was
saying that open source drivers would preclude the effectiveness of the
HFS. I do not believe this to be true, and this is also very different
from disallowing open source drivers altogether. However, upon carefully
reading it again I noticed that the author did assert that the
specification required that "the operational details of
the device be kept confidential". If this is true, it would cause
problems for Linux drivers. However, I cannot find the source for this
assumption, and the author merely states "a number of
the sources used were non-public". I am disinclined to believe such a
bold statement without any evidence. It would certainly be a drastic
change in the behaviour of companies like Intel to suddenly stop giving
out specs and releasing open source drivers.

I also think that the author is exaggerating many issues, and creating a
false impression of what will happen, probably because it seems that
many of the sources from Microsoft are exaggerating many issues and
creating a false impression of what will happen. Considering the speed
at which protections have been broken in the past, if drivers or devices
which are breached will have their signatures revoked as Microsoft
asserts, Vista users will have a very unpleasant experience. Microsoft
would destroy themselves if they were to revoke drivers and punish users
in the way that they describe. If such absurdly draconian measures were
enforced, consumers simply wouldn't buy premium content, and the market
for "non-participating devices" would be large enough that manufacturers
would continue to make them. However, I must say that I am disappointed
at the author's attempt to link this to "homeland security" and claims
that the protection will endanger lives, tactics which even Microsoft
did not use.

Most of this discussion probably belongs on sounder instead of
devel-discuss, unless I am mistaken in my understanding of the
devel/devel-discuss split, which seems to have caused considerable
confusion as the the proper lists.

Yours,
Constantine Evans