- From: Adam Roach <adam@nostrum.com>
- Date: Wed, 08 Oct 2014 17:31:45 -0500
- To: Chris Palmer <palmer@google.com>, Eric Rescorla <ekr@rtfm.com>
- CC: Anne van Kesteren <annevk@annevk.nl>, Justin Uberti <juberti@google.com>, Stefan HÃ¥kansson LK <stefan.lk.hakansson@ericsson.com>, "public-media-capture@w3.org" <public-media-capture@w3.org>
On 10/8/14 16:45, Chris Palmer wrote: > TL;DR: We don't have time, user attention, or space to communicate > crypto nuance. Therefore we must quantize the security guarantee > upward. Sure. You're getting off onto the tangent of opportunistic encryption, rather than really talking about the gUM issue. I'll post one quick rebuttal here, and then I intend to let the issue alone on this list (since it's several steps removed from the media capture charter). I think where you're misconstruing [1] what the proponents of opportunistic encryption [2] are proposing is that you're pretending that someone, somewhere has proposed that such connections should carry the full regalia of authenticated connections [3]. I don't believe anyone is seriously proposing that. I believe I'm speaking in alignment with most or all of the proponents of opportunistic encryption when I say that the expectation is that the user-visible interface would render such connections as "insecure." They just *happen* to be impervious to passive attacks. The user doesn't need to know about this to benefit; and I think we're in agreement that telling them about such a nuanced situation is likely to lead them to act in ways that are counter to their interest. The only real difference between deploying or not deploying opportunistic encryption is that deploying it makes things non-trivially better for users' privacy, even if they don't know about it. /a ____ [1] I'm giving you the benefit of the doubt here that this is a simple misunderstanding rather than feigning ignorance, although the strawmen you're using have the unfortunate appearance of intentional caricatures rather than honest representations of the opposing position. Your pedigree implies that you might know better. [2] And I do mean *encryption*, not some pseudo-crypto obfuscation as you imply -- do you really think that's what people are suggesting? [3] e.g., A lock icon, maybe some green or blue splashed on the UI somewhere, additional information pulled from from the server cert, etc.
Received on Wednesday, 8 October 2014 22:32:21 UTC