Laser & DRM [LWN.net]

Laser & DRM

Posted Oct 5, 2006 13:42 UTC (Thu) by mingo (subscriber, #31122) [Link]

In your example, in my opinion, the problem lies with willful violation of safety regulations, not with GPL v3.

The problem lies in the GPLv3: in essence it declares the "tool of DRM" evil, without leaving for circumstances. It does so by defining the DRM keys to be part of the "Source Code" (see Section 1 of the GPLv3) - and hence forbids the non-release of keys upon redistribution (they are defined into source code, and thus must be released).

Without leaving for circumstances (except a limited list of hand-carved exceptions of a currently known good uses of DRM) it's just plain impossible to correctly map all "good" and all "evil" uses of DRM in advance. The GPLv3 allows little for the licensor (or the judge) to weigh circumstances - it forces the decision of "good vs. evil" right in the license. Furthermore, in its public messaging, the FSF does not even allow for the /possibility/ of "good" DRM uses - DRM has been villified in its entirety.

This is what i loosely described in other threads as "the GPL now gets into the business of defining good and evil, and it should not do so".

To go back to your suggestion: even if there's a blatant violation of safety regulations, if the manufacturer is proven to /not/ have done everything technically possible to protect health, it can be found liable. (there are precedents for that) So a manufacturer, if it cannot use DRM in its laser, X-ray, radio or radar machine might just pick another OS, just to be on the safe side.

And even if it's not found liable, the manufacturer's engineers might have (gasp) conscience.

Laser & DRM

Posted Oct 5, 2006 14:32 UTC (Thu) by nim-nim (subscriber, #34454) [Link]

> To go back to your suggestion: even if there's a blatant violation of safety
> regulations, if the manufacturer is proven to /not/ have done everything
> technically possible to protect health, it can be found liable. (there are
> precedents for that) So a manufacturer, if it cannot use DRM in its laser,
> X-ray, radio or radar machine might just pick another OS, just to be on the
> safe side.

If the tamper-proofness of the device is so important he should go ROM. DRM is not an absolute protection either.

Laser & DRM

Posted Oct 5, 2006 14:38 UTC (Thu) by stijn (subscriber, #570) [Link]

Right now I have not yet seen convincing examples where DRM could be Good. In the proposed scenarios the Goodness comes from preventing certain evil things, where to me it seems that those evil things are illegal to begin with and can be prevented by other means, e.g. contract, warranty, ownership, access and probably a few others. The proposed scenarios, as far as I am aware, furthermore have so far been thought experiments. Correct me if I am wrong. As for the "engineer's conscience", we drive cars that are well capable of killing ourselves and other people with a minimum of effort (and the list is neverending), never mind guns.

So on the one hand we have evil DRM marching in on our lives, limiting our rights and taking away stewardship, with no Good specimen yet to be discovered. On the other hand we have the possibility that Good DRM may further our lives at some point in the futures. As far as I am aware, the Good DRM in those scenarios 1) can be replaced by other means and 2) is likely not a definite remedy to the problem it is supposed to solve either.

I appreciate your concern about 'defining good and evil', but the considerations above and the scenarios thus far have to my mind not yet advanced the DRM case.

Then there is the issue whether the DRM clause now makes v3 transcend software by tying keys to the source. The discussions elsewhere have not convinced me that Pandora's box has been opened. It is a view consistent with the four freedoms that Tivoization violates the spirit of GPL v2, although I am still trying to decipher your Tivo/PC/harddisk/binary example elsethread!

Laser & DRM

Posted Oct 5, 2006 19:29 UTC (Thu) by drag (subscriber, #31333) [Link]

Exactly.

It's like saying:

Hey barcode readers owners shouldn't be allowed to tweak their devices. Therefore the GPLv2 is bad because it doesn't allow people to embed binary-only drivers into the Linux kernel to prevent this!

Also manufacturers are legally obligated to restrict certain frequencies and power outputs on their Wifi cards. The Linux kernel being GPLv2 doesn't allow people to embed binary modules to prevent this and thus GPLv2 is bad.

It's realy kinda silly.