Conversation

Notices

1. Christine Lemmer-Webber (cwebber@octodon.social)'s status on Friday, 28-Jun-2019 18:15:35 UTC Christine Lemmer-Webber

   • Alexandre Oliva
   • Pablo Virgo

   @ptvirgo I suppose I should be clear by what I mean by "victim blaming", and that I do not think that @lxoliva is a "victim blame-r".

   In this particular case, the kind of vulnerability involved was a memory vulnerability. While this had to do with javascript execution, it could have as easily been based on image rendering.

   (cotd...)

   • Christine Lemmer-Webber repeated this.
   • Christine Lemmer-Webber (cwebber@octodon.social)'s status on Friday, 28-Jun-2019 18:17:22 UTC Christine Lemmer-Webber

     in reply to

     • Alexandre Oliva
     • Pablo Virgo

     @ptvirgo @lxoliva The response that "you should only run software you trust" then doesn't really apply to the general class of vulnerability that was exploited. If it was instead rendering an image that caused the overflow, the effect would be the same, and we *would* say anyway that visiting a webpage with firefox... well, firefox is a "trusted" piece of software as much as any large software can be trusted.

     (cotd ...)

     Christine Lemmer-Webber repeated this.
   • Christine Lemmer-Webber (cwebber@octodon.social)'s status on Friday, 28-Jun-2019 18:20:20 UTC Christine Lemmer-Webber

     in reply to

     • Alexandre Oliva
     • Libre Lounge
     • Pablo Virgo

     @ptvirgo @lxoliva User / computing freedom IMO involves your computer *doing the things you want and intend*. So that kind of vulnerability is not ok, and there *are* other architectural approaches that can protect users.

     Should we also only be running software out of a known and trusted free software corpus? Of course I think that! If you listen to the @librelounge on Guix you'll hear me arguing for this: https://librelounge.org/episodes/episode-23-guix-with-ludovic-court%C3%A8s.html

     But these are not mutually exclusive. (cotd ...)

   • Lars Wirzenius (liw@toot.liw.fi)'s status on Friday, 28-Jun-2019 18:21:25 UTC Lars Wirzenius

     in reply to

     • Alexandre Oliva
     • Pablo Virgo

     @cwebber @ptvirgo @lxoliva If I only ran software I trust, I wouldn't run any software, ever.

     Christine Lemmer-Webber repeated this.
   • Christine Lemmer-Webber (cwebber@octodon.social)'s status on Friday, 28-Jun-2019 18:24:24 UTC Christine Lemmer-Webber

     in reply to

     • Alexandre Oliva
     • Pablo Virgo

     @ptvirgo @lxoliva We should be advocating for both of these things in tandem: fully reproducible, free software, reproducible and auditable software.

     But for the reasons I explained earlier in the thread, *it isn't possible to trust all software anymore*, and not only on a "well it's free software thus it's secure" perspective either. We need to do better. We need *secure architectures* along with free software, so that damage can be limited.

     These. Are. Not. Mutually. Exclusive.

     Christine Lemmer-Webber repeated this.
   • Christine Lemmer-Webber (cwebber@octodon.social)'s status on Friday, 28-Jun-2019 18:37:13 UTC Christine Lemmer-Webber

     in reply to

     • Alexandre Oliva
     • Pablo Virgo

     @ptvirgo @lxoliva So here's what I mean by "victim blaming": looking at how someone was exploited in a way that, architecturally, is very hard to avoid today, and saying "oh, look, we can trace back the mistake they made" when instead we could take an architectural approach to make things safe by default.

     This is why I'm pushing for thing like ocaps, memory safety, more secure hardware. Nobody can trust all their software today, free or proprietary. It's just not possible anymore.