[liberationtech] Fwd: Re: PRISM and an Agenda for European Network

[Christian Grothoff]

On 07/05/13 13:51, Ralph Holz wrote:
> -------- Original Message --------
> Subject: Re: [liberationtech] PRISM and an Agenda for European Network
> Date: Fri, 05 Jul 2013 02:28:03 -0400
> From: Jonathan Wilkes<jancsika at yahoo.com>
> Reply-To: liberationtech<liberationtech at lists.stanford.edu>
> To: liberationtech at lists.stanford.edu
>
>> * GNUnet seeks to make decentralized systems: faster, more scalable;
>> easier to develop, deploy, and use;
>
> What exactly is GNUnet?
>
> To put it in gaming parlance (useful because games-- or at
> least the gaming community-- take UX seriously), is it
> currently the equivalent of a prototype for a
> gaming engine that has a GUI for the sake of convenience
> and development
>
> Or is it a prototype for a system that includes
> a GUI that an end-user would eventually use?
> (I.e., the "game" itself.)
>
> Or something else?
>
> You write that it aims to make decentralized system easier
> to use, but where does the user actually fit in to the picture?
>
> I sometimes feel like GNU forgets that "freedom to run the
> program for any purpose" is moot if merely running the
> program requires skills that 99% of the population doesn't have.
>
> -Jonathan
> --

Let me ignore your GNU-bashing.

As to your real question, as with any gaming prototype, you cannot 
expect 99% of the population to be able to use it while it is a 
prototype, so if you try now you'll likely be disappointed --- if you 
take the end-user perspective.

Now, to really answer your question, let me elaborate a bit.  In
GNUnet (specifically, gnunet-gtk) we do have graphical user
interfaces that we hope will eventually be useful for "real" users.

However, I'm not a GUI expert and while I agree that GUIs are
important, my personal focus is a bit lower in the stack. Now, this
said, the GNUnet architecture is designed to make it _EASY_ to write
GUIs for the system by essentially offering libraries (currently in C 
and Java) to access the system.  We usually begin by writing a geeky 
command-line tool for accessing and interacting with each subsystem
(the command-line tool links against the respective library for the
subsystem), and then (if the subsystem is high-level enough) eventually
get around to writing a graphical user interface.

So, if you don't like our GUIs or are an expert in usability, please do
come forward, tell us how we can improve or even better, send patches. 
Or write your own GUI from scratch.  Our APIs are very well documented 
(https://gnunet.org/svn/gnunet/src/include/), so if you think the 
biggest issue is GUIs for the 99%, please do help us fix that, I'm 
_always_ happy to see our GUIs improve and I do believe strongly that 
developers should focus on issues they care deeply about, and in my 
opinion we so far have too few hackers focusing on GUI aspects for 
GNUnet (you can find open tickets on https://gnunet.org/bugs/ under 
"gnunet-gtk" if you want to see some of the issues we're aware of; 
however, these only relate to subsystems that already have GUIs, 
subsystems that don't even have a GUI are not listed there).


Happy hacking!

Christian
p.s.: I'm not on the list, and I would suggest that further discussions 
about GNUnet development should happen on the gnunet-developer 
mailinglist or the GNUnet IRC channel.