On Wed, May 03, 2006 at 06:20:19PM +0200, Florent Daigni?re (NextGen$) wrote: > * Matthew Toseland <toad at amphibian.dyndns.org> [2006-05-03 17:06:03]: > > > How many SoC projects can we realistically mentor? We have 7 > > applications at present, and have set our target to 3. One of the > > applications is sufficiently whacky that I will need to read several > > papers on his site in order to evaluate it; all the others would be > > useful. > > > > Ian, will you be available over the summer to mentor people? Many people > > can assist students, even if they themselves are working for us under > > SoC; what you can't do is VOTE on proposals and also be a submitter of > > proposals. > > It's not clear : previous statements from google, including the FAQ now > say that it's not possible :/ that we should 'focus' on either beeing a > mentor OR a student. > > > I should be here for the vast majority of the summer, > > although there will definitely be shortish periods when I am not > > available at all. > > > > IMHO: > > - The FUQID replacement is important. > > see below > > > - The search engine is moderately useful. > > - JFKi is moderately useful. > > - Load balancing/limiting simulations and implementation is very > > important. Our current algorithm *mostly* works, but opens up a new > > class of attacks on direct peers, which probably will not be mitigated > > sufficiently by premix routing. > > I don't think it matches the "self-contained" requirement :) Moreover, > it involves a fair bit of understanding of how freenet works and possibly > the black magic behind the routing... Not a SoC project imho.
Hmmm. You think? We've been discussing it for some time, and I'm not convinced that there will need to be much deep magic as far as *routing* goes. > > > - Local messaging and file sharing is useful. > > Eternal debate : do we want to link freenet with tools possibly > incitating to do copyright infringment ? Even if we don't bundle it, > shall freenet be the mentoring organization of such a project ? Well, we definitely should have hooks whereby applications can talk to their direct peers, for example for bookmark sharing. > btw, according to Bombe, Freemule is on the road ^-^ Freemule and other tools that do not take into account local trust relationships aren't likely to work well. > > > - STUN and UP&P is useful. > > I'm not convinced :) > > > - The ant routing proposal; I have no idea, need to investigate his > > site. > > we can't see proposals ^-^ maybe we should incitate people to cross post > their applications on @tech ? -- Matthew J Toseland - toad at amphibian.dyndns.org Freenet Project Official Codemonkey - http://freenetproject.org/ ICTHUS - Nothing is impossible. Our Boss says so. -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: Digital signature URL: <https://emu.freenetproject.org/pipermail/tech/attachments/20060503/3dff9ff1/attachment.pgp>
