> Date: Thu, 20 Jun 2002 20:07:33 -0400 > From: Keith Moore <[EMAIL PROTECTED]> > Message-ID: <[EMAIL PROTECTED]> > > | I don't think it's reasonable to allow host implementations to reverse > | this default - this will break apps. > > This cannot possibly be true, as all addresses are temporary in the > current environment (all globals anyway). The only difference is in > their expected duration.
true, it might be more precise to talk about the expected duration of addresses; all addresses are at least potentially 'temporary' in some sense. however we're using the term 'temporary addresses' as it's used in RFC 3041, rather than in the more general sense of the term 'temporary' > So, to work at all, all apps always need to work with temporary addresses, > as that's all they're ever going to get. that's rather unrealistic; the current architecture is nowhere close to supporting the ability of apps to survive arbitrary address changes. > I like the current default of preferring "normal" addresses over 3041 addrs, > and I certainly see no problem with an application toggle to change that. > But nor do I see any problem with a per system toggle. Eg: on my laptop, > which is rarely turned on for more than 24 hours, hence no app survives > longer than that, a 3041 address (default timers) is exactly as stable as > a global address, and if I wanted, I see no reason that I shouldn't make > all apps default to using them (without having to recompile every one). I have less problem with a per-system toggle that the user can set than with allowing the vendor to change the default. but the vast majority of users won't understand the implications of such a toggle, and implementors of apps for which temporary addresses are appropriate will find it easy to enable them. by the time your laptop's OS is upgraded to support temporary addresses, chances are that the apps will support them also... so I suspect that the main effect of a user-settable per-system toggle would be to cause more apps to break with little or no improvement in privacy. that and I think the SHOULD clause already allows vendors to change the default behavior (or to allow users to do so) when they fully understand the implications of that behavior. so the draft does not prohibit a vendor from putting such a toggle on your laptop OS, just as long as they explain the implicatoins of that behavior to you. however I think it would generally be a foolish thing for vendors to do (at least commercial vendors) as it would increase their support costs while providing little additional functionality to users. Keith -------------------------------------------------------------------- IETF IPng Working Group Mailing List IPng Home Page: http://playground.sun.com/ipng FTP archive: ftp://playground.sun.com/pub/ipng Direct all administrative requests to [EMAIL PROTECTED] --------------------------------------------------------------------
