User:Mstone/Rainflow: Difference between revisions

Revision as of 20:15, 29 May 2009

Carrying on the tradition of naming software via puns, I'll lay down some thoughts on how to answer the question:

When should our software automatically do risky thing X?

People have been concerned with this question (and with variants and related questions) for some time:

<trac>5657</trac>, on spoofing-resistant update algorithms for de-isolated activities

the devil is truly in the details.

Consequently, here's an attempt at details.

@@ Line 1: / Line 1: @@
-Carrying on the tradition of naming [[Rainbow|software]] via puns, I'll lay down some thoughts here about software for, among other things, answering the question:
+Carrying on the tradition of naming [[Rainbow|software]] via puns, I'll lay down some thoughts on how to answer the question:
-: ''When should we run program X deisolated?''
+: ''When should our software automatically do risky thing X?''
 == Background ==
+People have been concerned with this question (and with variants and related questions) for some time:
-<trac>5657</trac> asked for a way to automatically update [[Terminal]] that is not subject to spoofing.
-One natural approach this is activity signing. However, as experience with X.509 has shown, the devil is truly in the details.
-To date, we have seen several attempts to discover appropriate details:
+: <trac>5657</trac>, on spoofing-resistant update algorithms for de-isolated activities
 : [http://lists.laptop.org/pipermail/security/2008-October/000496.html questions on activity signing and update thread]
 : [[User:Mstone/Commentaries/Bundles_1|activity semantics conversation]]
@@ Line 17: / Line 14: @@
 : [http://lists.laptop.org/pipermail/devel/2008-March/012131.html horizontal distribution thread]
 : [[Bundles and updates|homunq's ideas on bundles and updates]]
+the devil is truly in the details.
 Consequently, here's an attempt at details.