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

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

From: David Huynh <dfhuynh@alum.mit.edu>
Date: Fri, 22 Aug 2008 16:05:22 -0700
Message-ID: <48AF4632.20000@alum.mit.edu>
To: Dan Brickley <danbri@danbri.org>
CC: public-lod@w3.org, semantic-web@w3c.org

Dan Brickley wrote:
> David Huynh wrote:
> [snip]
> [many thanks for opensourcing btw!!]
>>> Maybe there's a chance to collaborate here, what do you think?
>>> I'm also not aware of any upcoming semweb UI workshops. WWW2009 
>>> might be
>>> a good place...
>> We can carry on this conversation off the list.
> Please don't dissapear offlist! At least www-archive@w3.org is always 
> available as a place you can have ad-hoc conversations be cc:'d to a 
> publically archived list. See 
> http://lists.w3.org/Archives/Public/www-archive/
> However we do already have a public-semweb-ui list,
> http://lists.w3.org/Archives/Public/public-semweb-ui/ ... where such 
> discussions are very much ontopic. That list deserves a bit of 
> attention and life. To sign up, do the usual W3C lists thing, ie. send 
> a mail to public-semweb-ui-request@w3.org with 'Subscribe' in the 
> subject line. Details at http://www.w3.org/Mail/Request
Ah, sorry, I was talking about Georgi's invitation to collaborate :-)

But let me take this chance to ask. What would people like to see in a 
paper to, say, WWW 2009, that talks about a UI paradigm for browsing 
graph-based data? One very obvious thing would be an overall 
implementation architecture that can be reproduced. Another would be 
some usage statistics. For example, UI event logs show that people did 
use all 4 one/many transitions

    many things to many things (pivot): 1088 events
    many things to one thing (focus): 878
    one thing to one thing (regular link): 1311
    one thing to many things: 450

Now these numbers don't prove anything in the rigorous sense, as the 
screencast itself might have influenced the users' behavior, but they 
might indicate something--not sure how to word that yet. One somewhat 
naive way to phrase this is that, if there is no way to move to a set of 
many things, then users are blocked (1088 + 450) / (1088 + 450 + 878 + 
1311) = 41% of the time.

There's also an idea to do a trivial pursuit game where users are given 
such a question as,
- "Were skyscrapers in Hong Kong designed by more Chinese architects or 
foreign architects?"
- "Which industries do large bank board members invest in? Give a 
distribution over the industries."
- "Are children of politicians likely to pursue careers in politics? Are 
they likely to succeed?"
We can measure success rate and learn something about how people 
translate such a question into a sequence of browsing actions.

Anyway, please let me know what you want to see in such a paper.

Received on Friday, 22 August 2008 23:06:11 UTC

This archive was generated by hypermail 2.4.0 : Thursday, 24 March 2022 20:41:12 UTC