Re: [Fenfire-dev] (no subject)

[Alatalo Toni]

> Toni, if you haven't already, you can join the mailing list here
> <http://interreality.org/lists>, and I won't have to approve future posts :)

i have but with a different (bad) address format..

> Here are some things that I though of while reading. Maybe I'll digest it
> further and post more later.  There are lots of things to think about!

we should somehow wrap this together soon enough (perhaps via a wiki,
coming up with common definitions etc?) to not explode the discussion and
get lost in it (on several lists)

seems you were not cc'ing vos-dev? i'm skipping that and Kyperjokki lists
here (too) but comment quickly (feel free to forward relevant parts to
lists if feel like it): (decided to copy fenfire-dev anyhow so that there
won't be duplicate answers (or if they are they supplement/correct/..
these))

> > that even though there are already many things in a place, or near some
> > object, there need not be basically any limit to the amount of
> > additional things that can be there, too.
> But how do you draw that? (Or should we *not* draw it?)  How do you write it
> down in a form that makes the links more immediate and clear than RDF?
> Is it possible?

there are several things, one is filtering -- using a subset of the
dimensions/relation-types for the rendering. e.g. in zigzag there may be
any number of dimensions (besides the default d.1 d.2 d.3), but most of
the views are capable of showing only 2 (2d) or 3 (pseudo-3d) at a time.

so you could say that the UI is actually 2d or 3d but there is an
underlying structure that has more.

dimensions i use often are d.clone (which connects the original cell to
all copies of it) and c.cellcreation (where (in gzigzag)) all cells
(nodes) are in the order of creation (kind of a time dimension).

and in the genealogy applitude you would have d.marriage and d.children
(with 3d perhaps something like d.date on z-axis)

the 'Compass' view in gzigzag shows connections in all dimensions with
names of the dimensions attached to the lines showing the connections, so
it is very similar to common visualisations of rdf graphs.

> Maybe, like ZigZag, we need ways for the user to play with the view of the
> structure: choose different mappings to the screen.  Is this hard to do in the
> general case?  Or are task-specific solutions better?

well, views to the general structure may be task-specific, like the
Presentation view I've often used with gzigzag instead of, say,
Powerpoint: http://an.org/gzz/screenshot-gzigzag-presentation.png
(that view is just a quick hack but still useful)

but there are a lot of things that may be general,
rows, columns, trees (horizontal/vertical) etc. come to mind,
gzigzag has these and some else implemented,
fenfire has had a bit different approach but the same could be ported
(e.g. to loom) quite straightforwardly i guess

> > http://an.org/gzz/screenshot-gzigzag-abc.png
> How usable is this? Do users "get" it?  What do you use to control the
> switching of views?

pressing 'v' (and shift-v for the other window) for next in list,
alt-v and alt-shift-v for previous, and the list is in the structure :)

i don't know much about other users than myself, except a bit about the
father of the idea Nelson

another pioneer, Douglas Engelbart, also emphasizes having different views
to the same structures (www.bootstrap.org seems to off-line now) and there
must be quite a bit of it in the hypermedia literature.

> > RDF
> > a-b-c
> > a-c-b
> I guess you could draw C and B twice, or try to draw all the edges.

yes, what Fenfire Loom seems to do is to draw C twice like this when
B is in focus: http://an.org/fenfire/screenshot-loom-abc_b.png
and B when C in focus: http://an.org/fenfire/screenshot-loom-abc_c.png

what kind of lacks currently is visualising the sameness of multiplaced
nodes, this was discussed in
http://mail.gnu.org/archive/html/fenfire-rdf-discuss/2003-03/msg00000.html

> > We present Storm,
> Storm is something like Freenet, where docs have unique id's that can be
> searched for?  I will have to check out that paper, sounds interesting.

yes, with an important distinction in the goals: Freenet aims at
anonymity, Storm at persistency. the idea is that users can store all
their data in Storm (which doesn't imply sharing, may just be a directory
in your filesystem where the Storm blocks are) and trust that it remains
accessible as long as it exists somewhere (e.g. in the p2p net).

> reed

~Toni