- From: Mark Nottingham <mnot@mnot.net>
- Date: Mon, 19 Jan 2015 16:53:29 +1100
- To: Henri Sivonen <hsivonen@hsivonen.fi>
- Cc: Chris Palmer <palmer@google.com>, Noah Mendelsohn <nrm@arcanedomain.com>, "Michael[tm] Smith" <mike@w3.org>, Tim Berners-Lee <timbl@w3.org>, Public TAG List <www-tag@w3.org>
Hey Henri, > On 16 Jan 2015, at 12:58 am, Henri Sivonen <hsivonen@hsivonen.fi> wrote: > > On Tue, Jan 13, 2015 at 8:59 PM, Chris Palmer <palmer@google.com> wrote: >> As discussed earlier in this thread, HTTPS requires clients to >> knowingly opt in to caching, transforming, or spying proxies. But such >> proxies are still possible. HTTPS makes them prove some value. > > While that's technically true, what you say assumes that users aren't > given an informed choice to make about the value. It's way too easy to > make up some excuse why the user needs to run an installer in order > for the Internet connection to "work" and such an installer could add > root certs so that browsers treat them as trusted root certs. > > I think the TAG finding shouldn't suggest that MITMing https might be > OK in some circumstances, because then those who want to MITM could be > emboldened to MITM and to claim that whatever they do is endorsed by > the W3C--all without giving users an opportunity to make an informed > choice and without actually matching the circumstances that the TAG > might have had in mind. The draft currently says: > Adopting "https://" has the side effect of disallowing shared HTTP caching [RFC7234]. Shared caching has a limited role on the Web today; many high traffic sites either discourage caching with metadata, or disallow it by already using "https://". However, shared caching is still considered desirable by some (e.g., in limited networks); in some cases, it might be so desirable that networks require users to accept TLS Man-in-the-Middle -- which is a bad outcome for Web security overall. Therefore, we encourage exploration of alternative mechanisms that preserve security more robustly, such as certain uses of Subresource Integrity [SRI]. Is that adequate, and if not, can you suggest edits? From my standpoint, I think we need to walk the line between acknowledging that when someone (e.g., your employer) owns the machine, they can and sometimes do MITM, while not condoning that as a desirable outcome. That's why the draft encourages work on browser extension APIs, improving PKI UX, etc. To the latter point -- I still find it remarkable that this is extremely common practice: http://ist.mit.edu/certificates ... and the OS/browser UX doesn't warn the user of the power granted by doing so (last I checked). Cheers, > For example, to stick with the place that > inspired TimBL's remarks, Great Britain is not really a *remote* > island for connectivity purposes, but it reportedly turned into a > massive captive portal lately > (http://arstechnica.com/tech-policy/2014/12/bt-sky-and-virgin-hijacking-browsers-to-push-porn-blocks/), > so there's a pressure to MITM without "remoteness" in the network > topology. > >> Overall, TBL seems to be saying that people shouldn't spy on the net, >> so that we can enjoy many social goods. Among those goods, he seems to >> place the ability to not have to adopt HTTPS. Unfortunately, we don't >> like in so innocent a world, and HTTPS is the bare minimum protection >> against tampering and spying. > > Yeah. The notion that https should be avoided on performance grounds > and the would-be snoopers be asked not to snoop seems unrealistic both > on the point of badness of the effect on performance and on the > effectiveness of just asking the would-be snoops not to. > > -- > Henri Sivonen > hsivonen@hsivonen.fi > https://hsivonen.fi/ -- Mark Nottingham https://www.mnot.net/
Received on Monday, 19 January 2015 05:53:58 UTC