On Sat, Nov 23, 2002 at 12:33:45PM -0800, Ian Clarke wrote: > I think we should get rid of the transient setting, the reason > being that it is relatively straightforward for a node to guess > whether it is likely to be transient or not. > > I would suggest that all nodes remain in "transient" mode > automatically until 6 hours after they come on-line. Nodes could > also keep track of what percentage of the day they are online > for, and switch back to transient if it is less than a certain > percentage (say 80%). > > This gets rid of the ugly manual transient setting. > > By a similar token, we should investigate whether the FNP > protocol can be enhanced to support a "tell me my IP address" > message, so that we have a reliable way to automatically > determine a nodes IP address. We can certainly autodetect IP for transients; the only downside is that we don't want to make transient more attractive to the user. Currently permanent nodes with non-internet addresses, or with things like localhost, are downgraded to transient and produce a warning on the mainport homepage. If that remains, I have no problem with trying to detect the real IP address and maybe downgrading to transient and grumbling while it is wrong (with a dyndns address, it might be wrong for a short while while it propagates... but you might argue that if the user has set a DNS address up they probably know what they are doing and it may not be worth checking)... maybe the best way to sort out this argument is for me to go implement ARKs :), I doubt it's a huge amount of work. It might be worth the further delay to 0.5.1 it would cause. It would allow us to completely avoid the setting ipAddress, cleanly, and we could then implement whatever heuristic to determine transience. The network may not be usable enough for it to be any use though (I do think we should consider raising the maxHTL a bit (35 say?) because of the network growth - but given the likely large number of pre-0.5.1 nodes out there, I'm not sure that we can...). I do agree with the principle that we should make it easier to deploy freenet nodes. > > Ian. > > -- > Ian Clarke ian@[freenetproject.org|locut.us|cematics.com] > Latest Project http://cematics.com/kanzi > Personal Homepage http://locut.us/
-- Matthew Toseland [EMAIL PROTECTED] [EMAIL PROTECTED] Freenet/Coldstore open source hacker. Employed full time by Freenet Project Inc. from 11/9/02 to 11/1/03 http://freenetproject.org/
msg05619/pgp00000.pgp
Description: PGP signature
