On Friday 17 Feb 2012 09:49:59 Florent Daigniere wrote: > On Thu, Feb 16, 2012 at 11:48:32PM +0000, Matthew Toseland wrote: > > On Thursday 16 Feb 2012 23:26:03 Florent Daigniere wrote: > > > > > > Different codebase, arguably no active developpement (7months since last > > > commit? - https://github.com/sanity/tahrir/commits/master), no release > > > yet, no userbase, ... different goals > > > > > > I am not sure it's a good fit for GSoC tbh... but then again, it's up to > > > Google to judge that. > > > Why do you think Freenet would be the right umbrella? Is there any > > > integration in between both planned on the roadmap? > > > > > > Why not presenting it as a mentoring organization on its own? > > > > Because setting up a nonprofit is a lot of work, and it's political enough > > that it'd need its own nonprofit, the standard ones might not be > > appropriate. > > > > That's a valid point I'm happy with :) > > > And because its goals are broadly similar ideologically even if its > > technical goals are, in the short term, somewhat divergent. And a lot of > > the technology is similar too, although of course he's done everything in > > New And Wonderful Ways. > > > > Agreed > > > Long term, Freenet needs pub/sub (for chat e.g.), and a good microblogging > > UI (well duh), and Tahrir needs routing and file storage (where are you > > gonna put your suppressed pictures/phone footage?). > > > > Yeah, I thought about that, hence I was asking about the roadmap. > > > Hence bringing it under the FPI organisational umbrella may make sense. > > Having said that, I don't think it would be appropriate to call Tahrir > > Freenet 2.0, at least not at present. :) > > You have convinced me :) > > My grumpy style probably suggested that I am against the idea... whereas I > was just trying to build my opinion on the matter. > > Florent
Another thing they both have in common: Spam-proof fully distributed keyword
searching is hard. Tahrir needs this because Twitter-style services need
keyword searches as well as Follow. Freenet needs this for chat. Neither has a
plan for how to achieve it. IMHO we probably need some new semantics to enable
this in the long run. It has major security and technical issues though - e.g.
do you expect the node to know the full web of trust? How do you prevent DoS's?
What if it doesn't fit in RAM? Should we search with a generic identity, and
then filter client side, and if so how would we create one? What about
minorities which the generic identity distrusts?
Having said that, the simplest thing would be something like FASD
({keywords},{identities}) - or even just multi-SSKs. So identity + keyword -> {
multiple matches, limit, means to access more, preferably without needing to
keep state }. So the simplest form is multi-SSKs. Might want them to be even
smaller, so maybe you get DoS issues just with verification - but this ought to
be resolvable.
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ Devl mailing list Devl@freenetproject.org http://freenetproject.org/cgi-bin/mailman/listinfo/devl
