W3C home > Mailing lists > Public > semantic-web@w3.org > August 2008

RE: freebase parallax: user interface for browsing graphs of data

From: Georgi Kobilarov <gkob@gmx.de>
Date: Fri, 15 Aug 2008 02:46:17 +0200
Message-ID: <180C011CD4FF654AB4B73A9A5AD7472C0A48AF@aristoteles.zuhause.lan>
To: "Kingsley Idehen" <kidehen@openlinksw.com>
Cc: <public-lod@w3.org>, <semantic-web@w3c.org>

Hi Kingsley,

> Is this an apropos or an insight addition?
> 
> My entire position about Linked Data solutions is centered on "User
> Interaction".  Note, I said: "Visualizations are good, but graph
> visualizations are not the sole keys to the treasures that reside
> within
> a Graph".

True.

Well, my point was that I got the impression that you've characterized
David's interface as a graph visualizer. Which is, imho, not correct.

And I'd like to highlight the distinction between visualizations and
interaction models. The core of every interface is the interaction
model. On top of that, there might be specific visualizations. 

The problem is that most graph UIs I know are based on the "one resource
at a time" interaction model. It's the interaction model of the current
web, where a resource is a web page, and users interact by looking at
one page (reading it), and browsing from one page to another. That
interaction model underlies all current linked data browsers (Tabulator,
Disco, etc.). Some UIs use a graph visualization on top, but the
interaction model remains the same. (As a side note, I think that
visualizing a data as a graph is useless. Why using a 2-dimensional
layout where both dimensions are undefined?) 

But the real magic happens when you change to interaction model, and
users can interact with multiple resources at a time. David has shown
some excellent examples. Exhibit provided a faceted filtering
interaction for graph data. And Parallax now demonstrates a solution to
graph browsing.


> Give users the  "option" to use their own cognitive skills to interact
> with the data (e.g. pivot, project, and traverse based on their own
> specific needs).

Give users as few options as possible, and design them in a way, that
users do not have to use their cognitive skills.

> Programmers cannot acquire the domain expertise
> of an alien domain.

They should. Or at least have an interaction designer supporting them.


Best,
Georgi

--
Georgi Kobilarov
Freie Universitšt Berlin
www.georgikobilarov.com

> -----Original Message-----
> From: Kingsley Idehen [mailto:kidehen@openlinksw.com]
> Sent: Thursday, August 14, 2008 2:25 PM
> To: Georgi Kobilarov
> Cc: public-lod@w3.org; semantic-web@w3c.org
> Subject: Re: freebase parallax: user interface for browsing graphs of
> data
> 
> Georgi Kobilarov wrote:
> > Kingsley,
> >
> >
> >> Visualizations are good, but graph visualizations are not the sole
> >>
> > keys
> >
> >> to the treasures that reside within a Graph.
> >>
> >
> > It's not about graph visualization. It's about user interaction for
> > graph-based data.
> >
> Georgi,
> 
> Is this an apropos or an insight addition?
> 
> My entire position about Linked Data solutions is centered on "User
> Interaction".  Note, I said: "Visualizations are good, but graph
> visualizations are not the sole keys to the treasures that reside
> within
> a Graph".
> 
> I regard "Data Access" as part of "User Interaction".   This is why
the
> OpenLink Data Explorer extension (for instance) simply adds the
ability
> to "View | Linked Data Sources" to its host browser.  All it is really
> doing is providing the user with a route to a beachead from which a
> query could be beamed (i.e. SPARQL + Full Text under the covers across
> the data sources in it's history).
> 
> Give users the  "option" to use their own cognitive skills to interact
> with the data (e.g. pivot, project, and traverse based on their own
> specific needs).
> 
> We are repeating a long history of not understanding intersection of
> "User Interaction" and "User Cognitive Skills" if we don't provide
> routes to the data behind subject projections.  This matter has been
> the
> same since advent of the IT era:
> 
> 1. Executive Information Systems (what used to be called EIS)
> 2. Report Writing Systems (Cognos, Crystal Repors, Business Objects,
> Brio,  Forest & Trees, I can go on, used and worked with all of them)
> 3. Dektop Productivity Tools (Access, MS Query, Paradox, and many
more)
> 
> "User Interaction" is a about a rich set of options that provide
> outlets
> for cognitive prowess. Programmers cannot acquire the domain expertise
> of an alien domain.
> 
> BTW - I also gave an "User Interaction & Linked Data" talk / demo [2]
> at
> the most recent Cambridge Semantic Web Gathering.
> 
> Links
> 
> 1. http://ode.openlinksw.com
> 2. http://tinyurl.com/5pvbcz  -- a presentation remixing presentations
> by TimBL and I from LDP (focus on slides 14 to 21 as this is what ODE
> is
> all about revealing to the user as a Linked Data interaction option)
> 
> 
> 
> Kingsley
> > Georgi
> >
> > --
> > Georgi Kobilarov
> > Freie Universitšt Berlin
> > www.georgikobilarov.com
> >
> >
> >
> >> -----Original Message-----
> >> From: public-lod-request@w3.org [mailto:public-lod-request@w3.org]
> On
> >> Behalf Of Kingsley Idehen
> >> Sent: Wednesday, August 13, 2008 8:43 PM
> >> To: Richard H. McCullough
> >> Cc: David Huynh; public-lod@w3.org; semantic-web@w3c.org
> >> Subject: Re: freebase parallax: user interface for browsing graphs
> of
> >> data
> >>
> >>
> >> Richard H. McCullough wrote:
> >>
> >>> Very nice!!!
> >>>
> >>> How is your knowledge base structured?  What language?
> >>>
> >>> Can I download your program to my computer?
> >>>
> >>> Dick McCullough
> >>> Ayn Rand do speak od mKR done;
> >>> mKE do enhance od Real Intelligence done;
> >>> knowledge := man do identify od existent done;
> >>> knowledge haspart proposition list;
> >>> http://mKRmKE.org/
> >>>
> >> David,
> >>
> >> For purpose of clarity and broader discourse (I know we've been
over
> >> this in private), what is the Linked Data and/or Semantic Web
> oriented
> >> value of this very cool visualization?
> >>
> >> Put differently, it would be nice if I could beam a query down the
> >>
> > data
> >
> >> graph exposed by you very nice visualization rather than being
> >>
> > confined
> >
> >> to the options presented by your application.
> >>
> >> At the very least, what's the harm in exposing the Freebase URLs in
> >> these Web Pages? If you do that, at the very least, other user
> agents
> >> can do stuff with the graphs (Linked Data) that you are
visualizing.
> >> The cost of this little tweak is extremely low and the upside
> >>
> > extremely
> >
> >> high.
> >>
> >> Cool stuff for sure, but I would like Parallax to be a nice Linked
> >>
> > Data
> >
> >> Web contribution also :-)  Freebase (basic) is part of the Web, but
> >> Parallax is sort of confining me to a Freebase enclave by not
> exposing
> >> URLs (where I currently see: javascript:{}).
> >>
> >>
> >> My fundamental argument remains this:
> >>
> >> Visualizations are good, but graph visualizations are not the sole
> >>
> > keys
> >
> >> to the treasures that reside within a Graph. Effective traversal
> (e.g.
> >> query beaming SPARQL or MQL) is also part of the puzzle, and it
> would
> >> be
> >> nice if we could always offer the visualization and the
> >> graph-query-beam-beachhead as part of a single Web information
> >>
> > resource
> >
> >> deliverable. There are always reasons why one or more humans (due
> >> inherent cognitive prowess) would seek to view the same data
> >> differently, no matter how compelling the initial visualization,
due
> >>
> > to
> >
> >> the fact that data visualizations are inherently subjective
> >> projections.
> >>
> >> Links:
> >>
> >> 1. http://tinyurl.com/5of2qu - Abraham Lincoln as Linked Data from
> >> Freebase  ( I get no triples with Parallax pages since the Freebase
> >> Data
> >> Sources aren't exposed)
> >>
> >> Note: The Freebase RDFization Cartridge (Wrapper, Scrapper etc.)
> will
> >> be
> >> better i.e., right now there are too many literal property values
> that
> >> should be URIs (this fix is coming). Ditto proper lookup driven
> >>
> > meshing
> >
> >> with DBpedia.
> >>
> >>
> >> Kingsley
> >>
> >>> ----- Original Message ----- From: "David Huynh"
> >>>
> >> <dfhuynh@alum.mit.edu>
> >>
> >>> To: <public-lod@w3.org>; <semantic-web@w3c.org>
> >>> Sent: Tuesday, August 12, 2008 3:11 PM
> >>> Subject: freebase parallax: user interface for browsing graphs of
> >>>
> >> data
> >>
> >>>
> >>>> Hi all,
> >>>>
> >>>> I've been exploring some user interface ideas for browsing graphs
> >>>>
> >> (of
> >>
> >>>> data in Freebase) in a user-friendly way, which I think might be
> >>>> applicable to some of the data that you have. The screencast
> should
> >>>> explain:
> >>>>
> >>>>    http://mqlx.com/~david/parallax/
> >>>>
> >>>> Please let me know what you think!
> >>>>
> >>>> Thank you.
> >>>>
> >>>> David
> >>>>
> >>>>
> >>>>
> >>>>
> >>>>
> >>>
> >>>
> >>>
> >> --
> >>
> >>
> >> Regards,
> >>
> >> Kingsley Idehen	      Weblog:
> >>
> > http://www.openlinksw.com/blog/~kidehen
> >
> >> President & CEO
> >> OpenLink Software     Web: http://www.openlinksw.com
> >>
> >>
> >>
> >>
> >
> >
> >
> 
> 
> --
> 
> 
> Regards,
> 
> Kingsley Idehen	      Weblog:
http://www.openlinksw.com/blog/~kidehen
> President & CEO
> OpenLink Software     Web: http://www.openlinksw.com
> 
> 
Received on Friday, 15 August 2008 00:48:02 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 21:45:24 GMT