Re: [unrev-II] Greetings

From: Alex Shapiro (alex@touchgraph.com)
Date: Mon Nov 26 2001 - 21:43:51 PST

  • Next message: Gil Regev: "Social Networks Analysis, was RE: [unrev-II] Greetings"

    Hi Ping. 

    Great to hear from you on this list.  I have previously seen your work on criticons, and the Gnutella visualization.  What I didn't realize was that both were your projects.  I am very happy that you have been involved in both, because the two nicely bound my own research area:  Graph Based Interfaces for categorization, and more interestingly, GBI's as the evolution of hypertext.

    You can see demo's of some of my projects at http://www.touchgraph.com The two main examples are right on the main page, and others can be seen by clicking the "news" tab.  The most interesting one here is the map of weblogs, demonstrating TG's potential to depict social networks.

    Thanks for the links, I have lots to say about all of them so read on for comments.

    At 05:34 AM 11/26/01 -0800, Ka-Ping Yee wrote:

    Here are some tidbits:

        - Graph visualization has been mentioned here.  I recently
          did a project for an info-viz class that ended up as a paper;
          it got mentioned on Slashdot (probably due to our choice of
          Gnutella as an example domain, rather than the graph layout
          and animation technique that is the real work in the paper).
          See: http://bailando.sims.berkeley.edu/infovis.html

    This is a beautiful project.  I especially like the illustrations of the radial layout technique that have the purple background. ... The purple is there only for show, right?  It still looks cool though.

    I think that a radial layout is the optimal technique for illustrating the distance that nodes are from a selected focus node.  Using this technique to illustrate the connections in the Gnutella network is perfect, because each user will probably be most interested in themselves, and their immediate neighborhood. 

    I really like the way the transitions are done too.  One comment though, the layout does not take into account "non-tree" edges right?  The marital-ties graph look pretty good, but the nice layout was "faked" (to some degree), or was it? (It seems to me that had the layout been automatic, the graphs could have been rearranged to avoid intersections).

        - IBIS has been mentioned a lot.  One of my dreams is to build
          an e-mail archiver that produces output like this:
          See: http://lfw.org/ping/criticons/

    Very nice.  I especially like the fact that it seems that the pretty structure on the left could be generated automatically by putting the [#] marks in a text based email.  This would also make it easy to exchange mail with non-criticons enabled applications.  Seems like such an application would not be too hard to implement either, a simple script should do it.  The only thing that prevents me from doing so, (besides that fact that I feer that it's not as simple as it sounds), is that it seems that post-level organisation is more important then intra-post organization.     

        - Facets, categories, and metadata were also mentioned.  Marti
          Hearst, my advisor, is also interested in visualization
          and information retrieval.  I'm the core developer on her
          research project, Flamenco, which is about faceted metadata
          and how to make it usable.  Our team has designed, and i've
          built, a series of prototypes for an image database browser.
          It's not ready for prime-time, but if you're curious (and
          gentle with it), you might want to have a look:
          See: http://bailando.sims.berkeley.edu/flamenco/
          (Choose "development system", then try "FrankenFacet",
          "FrankenMatrix", "NewSingleTree", or "Interface 4".)

    Wow.  This is definately the way to go (as opposed to yahoo-like hierarchies), especially for images.  I think that this project might be very applicable to the http://www.gettyimages.com/ database.  (Do a search, and click on an image, you get a list of keywords that apply to it).

    The best part of the interface is the category refinements that appear to the left in the FrankenFacet, and FrankenMatrix views.  Having found an initial first category that one is interested in, one is now presented with a greatly reduced list of options that can help to quickly narrow the search.

    I really can't say enough good things about faceted classification.  It makes so much sence to assign more then just one category per object as hierarchies force you to do.  The only trick is to figure out how to navigate the resulting structure.  Flamenco seems to have got a lot of it figured out.  I like the horizontal divisions for the current category.  Something else that I though was cool, was when I selected a category with 5000 items, I was prompted to choose another category because there were too many items to display.  This is exactly the way it should work.

    I just played with that Smithsonian exhibit, "Revealing Things".
    It's pretty cool.



    -- ?!ng

    The only way to go is up.  The problem so far has been that people have been slow to transition away from metaphors based on physical space.  Thus the hierarchies, that closely model a file cabinet, or a library. 

    "Cyber" space allows one to represent objects in multiple dimensions (like thinkmap's Smithsonian museum exhibit), and whether one likes it or not, such objects spring up by themselves anyway (Gnutella networks for instance).  The trick now is to design better interfaces for navigating this space.  So... things are only going to get cooler.

    --Alex
    Yahoo! Groups Sponsor

    Community email addresses:
      Post message: unrev-II@onelist.com
      Subscribe:    unrev-II-subscribe@onelist.com
      Unsubscribe:  unrev-II-unsubscribe@onelist.com
      List owner:   unrev-II-owner@onelist.com

    Shortcut URL to this page:
      http://www.onelist.com/community/unrev-II


    Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.



    This archive was generated by hypermail 2.0.0 : Mon Nov 26 2001 - 21:15:09 PST