On Thursday 07 October 2010 22:47:48 [email protected] wrote: > Hello all, > > So, I didn't have time to do a mockup, due to an other project. But, I put > some thoughts in the user experience, and I think that it's actually a > better way of doing things (instead of doing a mockup without discussing > first the ground concepts). > > So, here it is, if someone can put it somewhere on the wiki (message to the > wiki administrators : I lost my password to my account, and don't seem to > have put a valid email (or it isn't my account, but I'm pretty sure I'm the > only one going by the nickname Dieppe :p)). > > Anyway, here it is (it's not well-formatted) : > > == Freenet UX improvements proposal == > > The goal of this document is to cover the current User eXperience of > Freenet, and list all possible improvements, in a comprehensive manner. I'm > not sure there is anything new, but it settle some ground rules about what > UX we want to provide. > > This is, imo, a very important subject, which have been neglected for too > long. If we want to keep users, we have to provide a good experience. It > will take some time before we agree on what to do, on how to do it, and to > actually do it, but it's worth it. > > We will soon (?) release freenet 0.8, and this will obviously not be ready > in time. But we should feature it for 0.9, with at least the UI problem > fixed. Darknet usability problem can be fixed in version 0.10 (or 1.0) . > > We only discuss what should be done here. We will discuss how once we agree > on this one. > > You are welcome to edit it, correct it, ... > > === Introduction : concepts and definitions === > > Freenet means two different things: the network, and the software. What is > of interest here is the software, and how the user interact with the > network. > We won't discuss how the network works. > > The software is composed of several components: > > - the network management (network logic plus network communication (udp)), > - the different way of communicating with the software (http, telnet, fcp), > - the user interface (FProxy), > - the "plugins", Java applications enhancing Freenet core. > > The Freenet network provides two main different features: > > - a secure and anonymous way of communication, > - a secure and anonymous storage system, > > and is split in two: > > - a secure, harvesting-free darknet mode, or F2F, > - a quick-and-easy entrance mode, subject to harvesting, "insecure" by > design, called opennet mode. > > === Current problems === > > I - Nobody uses the Darknet > > Most users don't use the darknet mode, because they either have no friend > using Freenet, or because it's too complicated. > Users *should* really use Darknet, as it greatly improves security over > Opennet. > > II - The UI is complicated > > Most users complain about Freenet being complicated. While it's true to a > certain extent, I'm sure we have room for improvements. > FProxy contains too many things. FProxy contains things not related: > freesite, download, network status, friends management, forum discussion, > WoT management... > There is also a lot of information shown that is often irrelevant to basic > users. > > III - The configuration page is crowded > > A project like Freenet should be "easy" to configure, and come with > reasonable defaults. > Currently, the configuration page is confusing, since you can configure the > webUI, the plugins, > but also the network, and your security options. > > === Proposed improvements === > > We'll call "plugins" application, since that is what they are. They are > application, using the Freenet library to communicate (I think we should > split the current freenet.jar in libfreenet.jar, and freenet.jar, or > FProxy.jar, but users tend to be confused by FProxy). > > (Note that plugins like UPnP, JStun, should be integrated as options, not as > plugins) > > I - > > Proposing applications that really use the friends could be an incentive for > user to switch to darknet. > For instance, we could imagine a facebook (or more diaspora) like > application that is based on Freenet friends (in fact, I think we should > provide an official > social networking application based on Freenet). > > We already have friend to friend messages. They're not used, or marginally, > because they're not easy to use, and are a bit hidden. > > So, we should provide an application to manage friends: > - add/delete friends, > - send/receive messages (with optional freemail notification). > > It's the basics. If we can, we should provide a full-blown social networking > application based on Freenet (would require some times, but would be > beneficial, see the diaspora buzz (we should have done that before)). > > Also, we should integrate the addition of new Friends in application like > Freetalk, where people all use Freenet and are at least acquaintance (I > don't know if it is feasible in a secure enough manner though). This way, > users can slowly migrate from Opennet to Darknet. > > II and III - > > Each feature should be treated as a different application, like google, > gmail, google maps, etc... > > The network configuration should be decoupled from the software > configuration. Since we already have a tray icon for windows and mac(?), we > should provide one for unix/linux, and put the network configuration there. > Note that it can also be relevant to change the network configuration in > other application, like tweaking the bandwidth in the dl/ul application. > > Google presents its most used apps in a small menu bar, we should do the > same, like we do with the minimalistic theme (it would be nice if we could > rearrange the menu given the usage statistic of the user). The other > applications are grouped on a page, with logo and description. We should do > the same. > > Each application should have its own configuration page, its own interface > (it is however important that we have at least one default theme common to > all official Freenet apps, but we should make the theming individual, > because it reinforce the idea that it's just web app based on Freenet, and > not really a part of Freenet). FProxy should be an entrance point, nothing > more, like the google homepage is. Users generally are accustomed to web app > now, and freenet based applications are web app, they're just running > locally. > > Plus, it should provide an easy way of installing a new application. We > should however make a clear distinction between official and non-official > apps. > > We shouldn't let the user install an application from the web: if they don't > have a working freenet node, they won't be able to use the application. It's > another reason why we should have plugins like UPnP built in. > > === END === > > If you have any comments, suggestion, please don't hesitate. It's not the > first time we speak about this, but I hope it's a bit more detailed than > usually. Also, this time I'll try to keep my promises, and actually do > something. There is nothing that should prevent me from doing so this time. >
Key issues here: 1. Freenet has two social networks: WoT and Darknet. Both are important, both need to be really easy to use, and it's likely that both will have various different applications for any given identity. We need some visual and naming conventions to keep the two clearly separate. I have explained why I think WoT in particular is absolutely core in another mail. 2. Because opennet is hopelessly insecure, and because Freenet might well be used in Bad Places in the not too distant future, 1) we cannot enable opennet without informed consent, and 2) anything that increases the barrier to entry to darknet is bad. I strongly support your calling plugins "apps", because thanks to Apple and Google everyone knows what apps are. This was not true when we originally added plugin support. IMHO moving network config to external tools is unnecessary and probably a bad idea. You might be right about UPnP/JSTUN. They have been plugins because 1) some java versioning issues, now fixed, and 2) security issues, including talking to central servers in the case of JSTUN - which arguably isn't a reason not to bundle it, just not to turn it on without informed consent. Most of the rest makes sense.
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ Devl mailing list [email protected] http://freenetproject.org/cgi-bin/mailman/listinfo/devl
