Forum Posts Feed
Visual representation of data
< Next Topic | Back to topic list | Previous Topic >
Posted by quant
Sep 1, 2010 at 09:33 AM
I was blown away when I’ve seen this:
http://www.youtube.com/watch?v=I2d312_dXEs
I can’t wait when some PIM will incorporate it and I will be able to navigate “universe of my knowledge” :)
Posted by Daly de Gagne
Sep 1, 2010 at 03:05 PM
Quant, very interesting.
How would you assess practicality of such software? It reminded me of The Personal Brain - of which I am never quite sure how helpful the GUI is.
The most recent date I could find on the program pages was 2009 - is it still being developed, I wonder, and if so, when will it be marketed as a standalone or part of another program?
Daly
quant wrote:
>I was blown away when I’ve seen
>this:
>http://www.youtube.com/watch?v=I2d312_dXEs
>
>I can’t wait when some PIM
>will incorporate it and I will be able to navigate “universe of my knowledge” :)
Posted by Tom S.
Sep 1, 2010 at 06:07 PM
I thought it was very cool. Unfortunately, as it looks here, it might be a lot more cool than useful. That’s a might busy interface for practical usage. The amount of knowledge you could display looks impressive but I’m not sure how I’d use such a thing.
Tom S.
Posted by quant
Sep 1, 2010 at 08:26 PM
I think it’s not just the amount of data, it’s the “data with relationships” that it could display efficiently.
Once you start to have multiple parents and links between “far” items, it’s impossible to display it in tree structure.
Also programs like the brain, connectd text, mind raider have problem with that, because it’s in plane.
This adds another dimension, look at how many triangles and cycled graphs are there in that video, which means highly structured data, if implemented properly (limiting displayed data by closedness attribute etc), this could display sooooo much more structure that it could be a real eye opener - it could give you new ideas, because you’d suddenly see something that’s just not possible now. At 7:00 there is an example where how the graph and functions would allow you to see certain relationship between data (strict pecking order), that might be difficult to figure otherwise or maybe you wouldn’t even think there could be some.
The colors, links, directions, weight, everything is programmable, it could have soo many functionalities implemented ...
Posted by Manfred
Sep 1, 2010 at 10:23 PM
For what it’s worth, ConnectedText’s “graph command” lets you represent the data as “non directed” (or non-acyclic) graphs that do show the relationships and how close they are. The graph command is different from the Navigator, by the way.
You can specify the depth of the graph, i.e. how many levels deep you want to go. My experience is that in a large project with many interconnections, more than 3 levels is too deep for any graph to make sense.
Still, you can show how close or distant a topic is from another topic and what interconnections exist between different topics. It’s quite useful.
I agree that a tree cannot deal with this. You need a more flexible data structure. But it is possible to do a lot on a plane. One does not necessarily need 3 D, unless one wants to go to higher levels of complexity than two or three levels deep. (It seems to me that if computer screens were larger, even a depth of four or five levels would make sense without 3 D.
Manfred