--- Log opened Sat Nov 26 13:44:46 2016
13:44 -!- rah [rah@verain.settrans.net] has joined #copperhead
13:44 -!- Irssi: #copperhead: Total of 62 nicks [2 ops, 0 halfops, 0 voices, 60 normal]
13:44 -!- Irssi: Join to #copperhead was synced in 0 secs
13:44 < rah> https://twitter.com/CopperheadOS/status/802158830889078789
13:45 < rah> what is the firmware this post is talking about?
13:45 < rah> which has "open sources" but isn't free software?
13:46 <@strcat> Pixel C boot chain is open source
13:46 <@strcat> it's signed though, ofc
13:47 <@strcat> so you can see the sources, and build it
13:47 <@strcat> but you could only flash a build signed by Google
13:47 <@strcat> kocka: it's not a typo
13:48 <@strcat> kocka: building docs are covering the current state in dev branch
13:48 < rah> I see
13:48 <@strcat> kocka: but 53 -> 63 isn't a security update
13:48 <@strcat> so there's no rush
13:48 <@strcat> beta channel updates really quickly
13:49 < rah> strcat: have the FSF commented on that?
13:50 < rah> I don't understand the criticism of the FSF
13:50 < rah> the FSF's position is that users should be free to do their computing how they wish
13:50 <@strcat> rah: they have a position on firmware signing
13:50 <@strcat> they're against it
13:50 <@strcat> they consider this evil: firmware can be updated, and there's signature verification
13:51 <@strcat> they don't consider this evil: firmware could be update, and there's signature verification, but the update mechanism was disabled in hardware at the factory
13:51 < rah> strcat: I don't think they're against firmware signing in general, just where the firmware signing prevents users from running the user's software
13:51 <@strcat> they are against it in general, since that's what it does
13:52 <@strcat> you can't have verified boot with a user controlled key, if earlier firmware is not verified against a hardware controlled key
13:53 <@strcat> so they are against verified boot unless the earlier boot chain cannot be updated... which would be bad too
13:53 < rah> as I understand it, the FSF's position is that hardware which the user cannot update is out of scope, not necessarily not evil
13:53 <@strcat> rah: no they're fine with it
13:53 <@strcat> the part they have an issue with is that the vendor can make an update, but the user can't
13:54 < rah> strcat: what do you mean by "fine with it" exactly?
13:54 <@strcat> if neither can, but the firmware is still there, just with update mechanism disabled, it's fine
13:54 <@strcat> rah: they don't consider it evil / they are happy to consider a free OS running on something with proprietary firmware okay if it can't be updated
13:54 <@strcat> their problem with it is not that it's there, but that the vendor can update it but the user can't
13:55 <@strcat> so if the update mechanism is disabled in hw, they're okay with it
13:55 < rah> strcat: can you provide a reference to a statement from the FSF that they do not consider it evil?
13:55 <@strcat> you can find it
13:56 < rah> I don't believe it exists
13:56 < rah> the FSF generally doesn't talk about things in terms of "evil"
13:56 < rah> they use words like "unethical"
13:56 <@strcat> rah: what's the difference
13:56 <@strcat> I'm not quoting anything
13:57 <@strcat> are you just going to switch to arguing about which specific word was used
13:57 < rah> switch from what? :-)
13:57 < rah> I don't think you understand what you're talking about
13:58 < rah> the FSF won't say "this firmware-which-can't-be-updated is not evil"
13:58 <@strcat> yeah, they will, because they consider it hardware, and while they like open hardware they don't consider non-open-hardware unethical
13:59 < rah> again, I don't think you understand what I'm saying
13:59 <@strcat> their problem with it is that the vendor can do something that the user can't
13:59 < rah> the FSF won't use words like that
13:59 < rah> the scope for the FSF is software
13:59 < rah> not hardware
14:00 <@strcat> they do make statements about whether hardware needs to be open, and how important it is relative to software
14:00 < rah> the only issue that the FSF is concerned about is: does the user have the right to do their computing as they wish?
14:00 <@strcat> so why is it okay if the vendor cannot make a firmware update, but not if the vendor can
14:00 <@strcat> if in both cases the user lacks control just as much
14:01 <@strcat> it's not about some level of user control / freedom in that case, it's the fact that the vendor has power that the user doesn't that they take issue with
14:01 < rah> if they vendor can update the firmware but the user can't then the vendor has the power to violate the user's ability to do their computing
14:01 < rah> s/if they/if the/
14:01 <@strcat> how can the vendor violate the user's ability to do their computing
14:02 <@strcat> the user can choose not to install their updates
14:02 <@strcat> I am not talking about an automatic update mechanism
14:02 <@strcat> only firmware signing / updates
14:03 <@strcat> 2 ARM development boards: 1 has boot chain that can be updated, 1 has the update mechanism disabled, and it has signing - same hw, they just burn some fuse on the board for the latter
14:03 <@strcat> and they release updates for the firmware
14:03 <@strcat> so users with the former can choose to install them, users with the latter cannot
14:03 <@strcat> FSF considers the former unethical, and the latter okay
14:04 <@strcat> where is the extra user control
14:04 < rah> this is not the case
14:04 <@strcat> it is the case
14:04 < rah> let me explain it to you
14:05 < rah> same 2 ARM development boards
14:05 < rah> they release updates for the firmware
14:05 <@strcat> but one has the update mechanism disabled in hardware
14:05 <@strcat> which makes it ethical per FSF
14:06 < rah> user downloads the source code for the firmware
14:06 < rah> user modifies source code to do what they want
14:06 <@strcat> and it has signing so they can't flash it
14:06 < rah> user installs firmware, machine won't boot
14:07 <@strcat> it won't install
14:07 <@strcat> one b/c signing fails, one b/c there's a burned fuse or w/e
14:07 < rah> well, it's not the same situation
14:08 <@strcat> it's not the same because the one with updates can at least be secured when vulnerabilities happen
14:08 <@strcat> and nothing prevents having open sources + reproducible builds despite signing
14:08 < rah> the firmware that they release is only for one board
14:09 < rah> there can never be any firmware released for the board with the burned fuse
14:09 <@strcat> rah: it can be for both and you just have to buy a new one to get it for the other
14:09 < rah> it can't be for both
14:09 < rah> the burned fuse board can never be updated
14:09 <@strcat> sure but a new one at the factory can have it
14:09 < rah> lol
14:09 <@strcat> vulnerability is found -> the burned fuse board has an unfixable security hole
14:10 <@strcat> you need a new one, if that matters for your use
14:10 < rah> what you're saying doesn't contradict what I've said
14:10 <@strcat> what good is verified boot if as soon as an openssl bug is found, there's an unfixable hole in the earlier boot chain
14:11 < rah> there can never be a firmware released for the board with the burned fuse
14:11 < rah> verified boot is not a requirement for user freedom
14:11 <@strcat> it is a requirement for a reasonably secure platform though
14:11 <@strcat> so I guess user freedom is incompatible with that
14:12 < rah> 07:24 < PaulFertser> "open sources + reproducible builds + signed updates" is better than proprietary indeed. But why should FSF be pleased if it's not free software? I think if we ask those in power there they all will agree. But it's not the point. FSF is promoting free software, the one that gives users power. CopperheadOS doesn't give users more power but it does make them more secure. Good thing. But unrelated."
14:12 <@strcat> that definition of user freedom
14:12 < rah> from #openmoko-cdevel
14:13 <@strcat> anyway GPL is proprietary, it hinders freedom compared to BSD/MIT
14:14 <@strcat> that's the FSF position, but moved over slightly
14:14 < rah> rofl
14:14 <@strcat> restrictions are okay, as long as it's their restrictions
14:14 <@strcat> everything is not black and white
14:14 < rah> indeed
14:15 <@strcat> if ppl are going to keep commenting that CopperheadOS is 'unethical' now because non-commercial use is no longer allowed without paying, after 2 years as FOSS, which didn't work out
14:15 <@strcat> then I'm going to keep bashing that ideology
14:15 <@strcat> FOSS CopperheadOS existed
14:15 <@strcat> for 2 years
14:15 -!- rah [rah@verain.settrans.net] has left #copperhead []
--- Log closed Sat Nov 26 14:15:35 2016