On Saturday 14 November 2009 16:47:27 [email protected] wrote:
> Great ideas from Toad and you.
> 
> But there is one big problem: we are quite late in the development
> cycle. We wanted to
> finish a working version till Christmas. Your new ideas would mean to
> make a big step backward
> and start another design from scratch (however, there isn't no design
> yet for the new ideas, only problems ;) ).

I'm not sure that is true. The mentioned code duplication in Freetalk hasn't 
been implemented yet afaik, and there are at least two WoT-based plugins making 
rapid progress other than Freetalk, hence the issue coming up in the first 
place.
> 
> When we containue as planned we wil release something that is surely
> not compatible with the new ideas.
> And clients will use the new interface and implement it the way you don't 
> want.
> 
> So the question is: start again with new design, or complete something
> that works soon?
> Or is there a way for both?

I am not sure that there is a fundamental problem here, isn't this mostly a 
matter of interfaces and of stuff that hasn't been implemented yet such as the 
Freetalk UI to WoT?
> 
> On Sat, Nov 14, 2009 at 17:36, Evan Daniel <[email protected]> wrote:
> > On Sat, Nov 14, 2009 at 10:08 AM, Matthew Toseland
> > <[email protected]> wrote:
> >> Currently Freetalk duplicates - or is planned to duplicate in the near 
> >> future - most of the functionality of the Web of Trust plugin e.g. it will 
> >> have identity pages with message counts etc. The plan is/was to hide the 
> >> Web of Trust plugin and just have everything under the Discussion menu.
> >>
> >> However, there are two additional plugins under heavy development which 
> >> use WoT:
> >> - digger3's WoT-based IRC system. This uses WoT for spam-resistant 
> >> discovery.
> >> - Artefact2's FlogHelper blog tool, which uses an identity from WoT to 
> >> simplify key management and eventually to avoid the need to announce the 
> >> site and to tie in with Freetalk for feedback.
> >>
> >> Plus, hopefully:
> >> - infinity0's distributed searching functionality in Library.
> >> - evanbd's Fritter microblogging app.
> >> - Private messaging functionality in Freetalk or a new WoT-based Freemail 
> >> version.
> >
> > So far I've been successful at slowly convincing digger3 to implement
> > most of the protocol-level ideas in my Fritter draft spec.  I'm
> > hopeful that the two overlap enough that (eventually) the only
> > difference between Freenet-IRC and Fritter would be the interface.
> > Conceptually, this is fairly easy: both are a collection of short
> > messages posted by users, and both need to solve the problem of low
> > latency messages without insane amounts of polling.  The same
> > related-links structure from Fritter works well with IRC.  IMHO you
> > can treat IRC as simply a different interface to a subset of the
> > Fritter functionality: IRC messages must have exactly one hashtag,
> > which is the name of the IRC channel.  (Then there's all the channel
> > moderator stuff like banning and topics, but I suspect that gets done
> > in a decentralized fashion that amounts to individual clients paying
> > attention to what the moderator publishes.)
> >
> > So I'm hopeful that digger3 will implement most or all of the Fritter
> > functionality while I continue trying to figure out how to take
> > meaningful statistics, and that at most it would just be a few changes
> > to the IRC plugin.
> >
> >>
> >> And probably more in future.
> >>
> >> Two basic problems here:
> >> 1. Code duplication: All these apps will need to duplicate much of 
> >> Freetalk's duplicated WoT stuff (e.g. nagging the user if their identity 
> >> hasn't been announced, setting trust etc). Most of them don't gain much 
> >> benefit from this.
> >> 2. The same identities will be reused for multiple applications, and it 
> >> should be easy to go from one to the next.
> >
> > I'm against code duplication, and I'm strongly against have different
> > UIs that do the same thing -- especially if changing something in one
> > UI means it changes in the other!  I'm also in favor of reusing
> > identities -- publishing a flog under the same SSK I use to insert
> > Freetalk messages and IRC messages is a good thing.
> >
> > Note that this brings up the past discussion of trust contexts.  My
> > recollection of that is we decided there was no reliable way to apply
> > trust ratings outside of the context they came from (when they come
> > from other people, that is -- if I trust / distrust a person locally,
> > it probably does make sense to apply that to all contexts).  Any
> > unified UI would need to handle this properly.  (And no, I don't have
> > any good suggestions for how.  I'm also skeptical that it can be
> > meaningfully simplified and retain its usefulness.)
> >
> >>
> >> IMHO the solution to both problems is to keep the Web of Trust plugin 
> >> visible, and make it easy to use. We might want to rename it ("Anonymous 
> >> Friends"? Any other ideas?). It is essentially an anonymous social 
> >> networking system: Each user has 1) various applications, and 2) trust 
> >> relationships with other users.
> >
> > I dislike overloading the word "friend" between darknet peers and
> > on-Freenet identities.  I propose "contacts" for the latter.
> >
> >>
> >> So, the profile page for a Known Identity should have links (and possibly 
> >> detailed information) for each of the user's contexts. So for Freetalk, 
> >> there would be how many messages he has posted, possibly a list of recent 
> >> messages and/or commonly posted to boards, and/or a link to a page in 
> >> Freetalk containing such information. For FlogHelper, there would be a 
> >> link to the identity's one and only flog. For private messaging, there 
> >> would be the ability to send a private message (whether by a link or 
> >> inline). And so on.
> >>
> >> We would keep the existing trust setting functionality, and we would have 
> >> a similar page for each of our Own Identities. In fact, we would probably 
> >> only need one menu item for Web of Trust, which would be a merger of Own 
> >> Identities with Known Identities: A list of Your Anonymous Identities, 
> >> click on one to go to its profile page, with the option to add more.
> >>
> >> The big gain is that you can find a message on Freetalk, click on its 
> >> author and send him a private message, change his trust levels, read his 
> >> flog, chat with him in real time, search or browse his published files 
> >> etc. This makes the Web of Trust plugin both useful and easy to use.
> >
> > I agree, this would be excellent!
> >
> >>
> >> We should be able to set the trust level from the own identity which is 
> >> logged in to the identity whose profile page we are looking at *on that 
> >> page*. We should probably also have some indication of how much we trust 
> >> the other identities listed on that page: although this complicates the 
> >> UI, it is very important practically IMHO.
> >>
> >> As p0s has rightly pointed out, when we go from Freetalk to a specific 
> >> identity's page, our main concern is him in the context of Freetalk: We 
> >> need to be able to see how many messages he has posted, and how many 
> >> messages those who trust him have posted, as well as their trust levels 
> >> etc. The current plan is to implement this page in Freetalk, fetch the 
> >> data we need from WoT, add what we know from Freetalk, and show a 
> >> Freetalk-focused page on that user. We can continue to do this, we just 
> >> need to fetch a little more from WoT: The links to the other apps. However 
> >> I expect that most WoT-based plugins will just use the WoT identity page.
> >
> > Click the name, get a freetalk-specific page that also has a link to
> > "view more about this contact" or something that shows the full WoT
> > page?  That would also provide other application-specific links.
> >
> > Evan Daniel

Attachment: signature.asc
Description: This is a digitally signed message part.

_______________________________________________
Devl mailing list
[email protected]
http://emu.freenetproject.org/cgi-bin/mailman/listinfo/devl

Reply via email to