[00:45:52] <toad_> okay, i have transcribed the various proposals on 
http://new-wiki.freenetproject.org/Roadmap/0.8.5 for darknet enhancements into 
a web of bugs starting at https://bugs.freenetproject.org/view.php?id=4621
[00:46:03] <toad_> I have absolutely no intention of doing all of them prior to 
0.8.0
[00:46:15] <toad_> what i consider to be vital is:
[00:46:22] <toad_> https://bugs.freenetproject.org/view.php?id=3919 fix Pitch 
Black attack
[00:46:33] <toad_> https://bugs.freenetproject.org/view.php?id=4622 per-peer 
trust level (get rid of friends seclevel)
[00:46:42] <toad_> https://bugs.freenetproject.org/view.php?id=4623 friend 
visibility setting
[00:46:49] <toad_> https://bugs.freenetproject.org/view.php?id=4624 connect to 
FOAFs
[00:47:01] <toad_> https://bugs.freenetproject.org/view.php?id=4625 routable 
connections to a subset of FOAFs
[00:47:15] <toad_> https://bugs.freenetproject.org/view.php?id=2604 use FOAF 
connections for connectivity purposes
[00:47:53] <toad_> https://bugs.freenetproject.org/view.php?id=4629 easy to use 
UI for upgrading a friend-of-a-friend to a full friend, possibly with out of 
band password verification
[00:48:06] <toad_> https://bugs.freenetproject.org/view.php?id=166 proper 
invite mechanism
[00:48:28] <toad_> https://bugs.freenetproject.org/view.php?id=4626 use friends 
in invites
[00:49:15] <toad_> https://bugs.freenetproject.org/view.php?id=1342 include the 
installer(s) in the invite (note that this may require some debugging - my node 
doesn't have the installers)
[00:49:42] <toad_> and one or other of:
[00:50:09] <toad_> https://bugs.freenetproject.org/view.php?id=4630 announce 
new FOAFs in case friends know them
[00:50:43] <toad_> https://bugs.freenetproject.org/view.php?id=4627 recommend a 
newbie to friends when creating an invite
[00:51:20] <toad_> https://bugs.freenetproject.org/view.php?id=1372 "Ian: Toad 
thinks you know Oskar"
[00:51:42] <toad_> fortunately, most of the above are fairly straightforward 
... some of them are not
[00:52:10] <toad_> we probably won't have a full solution to Pitch Black any 
time soon, i will implement oskar's proposal but there are some doubts about it
[00:52:37] <toad_> connect to FOAFs and invites are the core issues
[00:53:45] <toad_> I don't think bookmark sharing is essential, it'd be nice 
though
[00:53:59] <toad_> same for sharing completed downloads, making f2f file 
transfers work really well
[00:54:17] <toad_> although making f2f file transfers persistent (and therefore 
reliable) is tempting:
[00:54:32] <toad_> https://bugs.freenetproject.org/view.php?id=1461

4623 is trivial, modulo back compatibility code (an alert saying set them).
4622 is easy, has back compat issues, affects everything that uses the friend 
seclevel, but that's not much.
3919 is as noted significant - oskar's basic fix is easy, the simulation might 
take a couple of days, doing a proper set of attack simulations and getting 
something that's completely solid could take a lot longer and I'm not going to 
do it any time soon; maybe Evan or someone will later.
4624 is a few days work, hopefully
4625 should be easy once 4624 is done
2604 is probably a few days work
4629 including backend and frontend is probably a few days work
166 i'm not sure about, probably a few days to a week including 4626 and 1342; 
debugging related to 1342 shouldn't add much more as it's probably a priority 
issue
4630 is fairly trivial given the previous, less than a day
4627 and/or 1372 might be a day or two including the mechanism and the UI
1461 could be two days

(I am omitting pretty much all of the functionality enhancements here; this is 
stuff related to making darknet itself easy to use, easy to bootstrap, and 
perform well, the functionality enhancements, like bookmark sharing, sharing 
your downloads - all optional and related to friend trust level - can wait).

Which adds up to about a month, for a plausible estimate (not too optimistic or 
too pessimistic). However, we also have to get new load management working, 
which in the plausible worst case could take a month by itself (although there 
is some time before that that doesn't count due to not wanting to deploy 
*everything* at once).

We also have to consider that we haven't had a release for ages, that Freenet 
is needed more than ever and publicity would probably go down pretty well if 
it's in the fairly near future. Realistically I expect at least a couple months 
debugging after the feature-complete alpha. I did offer Ian a deadline of 31 
January; in light of the above I doubt this is feasible, and I strongly object 
to releasing without most of the above (because we don't want to get people 
onto opennet who really need protection, and the new opennet wording is a 
little more conservative/clear that opennet sucks), and I equally strongly 
object to releasing without finishing new-load-management (for performance 
reasons and because it is nearly there and because it helps on darknet and 
because the process of merging it has been somewhat chaotic; if we abandon it 
we'd need at a minimum to reinstate fairness between request types and have a 
stabilisation period).
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 197 bytes
Desc: This is a digitally signed message part.
URL: 
<https://emu.freenetproject.org/pipermail/devl/attachments/20101224/be8d5034/attachment.pgp>

Reply via email to