- From: Anders Rundgren <anders.rundgren.net@gmail.com>
- Date: Tue, 11 Nov 2014 15:42:05 +0100
- To: "public-web-security@w3.org" <public-web-security@w3.org>
For structuring the discussions a bit... It seems that there are three different user trust decisions that could apply to WebCrypto.Next: 1. Giving one of your identities to a site. This is already an established use-case on the web and I guess we won't be able to get away from such decisions. 2. During installation of a software package running with extended privileges, (hopefully) received through some trustworthy channel. Although I personally do not believe that signed web-applications is ever going to be mainstream except on "WebOSes", some people feel different. Anyway, it is a valid use-case since we already do this for native applications. 3. NEW. Some of the thoughts that have been floating around require the user to give site-specific code direct access to for example WebCrypto or PKCS #11 methods. IMO, this is an invalid use-case because it 1) is too fuzzy for users, 2) doesn't match typical issuer policies, 3) introduces considerable vulnerabilities. AndersR
Received on Tuesday, 11 November 2014 14:42:37 UTC