Re: Yarrr UI

[Seth Nickell <seth gnome org>]

I think the big thing is we don't know how this will work. Some things
will work worse, there's absolutely no question. The first UI approach
here is to get at the "meat" of the idea... to get it in its most pure
form. Some things will probably work, and some things will crash and
burn. We'll find we'll need to meld in certain traditional bits (I think
its likely we'll end up w/ a simple way to just create a comment
directly w/o the chat, for example, but I want to try the "pure way"
first).

Our primary goal right now is to do a very clean implementation of the
concept, and then try and coax people into using it. That will get us a
lot more information about what actually works, and what doesn't. Social
software is much harder to predict and design perfectly in advance (e.g.
there's no way Joel could have done that stuff without seeing all the
train wrecks of comments systems before him that failed... and observing
and thinking how).

> [ This actually is one of my doubts about the design. One property of
> IRC discussions is they are mostly untargeted, mixed and not specific,
> which is part of the fun of the medium. Will the same work when applied
> to discussion and cooperation on a specific topic? Or is a delayed
> discussion, where you have more time to think and to express concepts,
> appropriate?]

A different variation of the design I played with had a single
discussion area at the top of the topic. Then you'd select different
parts of the conversation and "summarize" it. Alternatively, you could
just lose the conversational context and not worry which comments tied
into which part of chat (not many people will read it anyway, its not
worth changing the UI around too much).

Anyway, I think it is clearly important to have "social" "bullshitting"
sorts of conversations occuring somehow. Its clearly part of what makes
IRC work, and if we don't provide a way, nobody will use it :-) Of
course, having it more focused than IRC will also hopefully drop some of
the noise and encourage chat to be used for actually resolving issues
(as opposed to, say, #gnome-hackers which is just a cespool of social
conversations today).

> Something obvious is that the one line entry is targeted to interactive
> chat. Also would a mixed interactive/delayed discussion be readable?

I'm not sure. In general, its very annoying to read through IRC logs
that aren't currently live discussions. I cringe when somebody sends me
a log and says "We discussed such and such, can you read this and tell
me what you think" ? :-) We'll just have to see how it goes! 

> Anyway, I think this is a good base to experiment on. I think some
> specific aspects will work, while some others will not. It's where we
> should use an experimental approach rather than a design one, though.

Yup! I agree totally, Some stuff will be totally wrong, but hopefully
we'll find some interesting ideas. By trying this in a "pure"
experimental form rather than including lots of potentially useful stuff
(file attachments, etc) we can get a better idea of what works and what
doesn't without a lot of distraction.

-Seth