3D Computer Network Maps

beebware writes: "According to this article on C|Net, Tim Bray (co-inventor of XML) has launched Antarcti.ca which renders computer networks in 2 and 3D maps. It's currently running a demo off the ODP data. But will it take off? Will users really like 'country-maps' opposed to listings? (Incidentally Tim used to be vice-president of production at Yahoo! so I think we can tell what his money's on.)"

Comments

Correction: Bray has no connection to Yahoo

[Ars-Fartsica]

There is a former VP of production (promoted, now has another title) at Yahoo who has a name very, very similar to Tim Bray, but they are not the same person.

Good Visualization Enhances Good Data

[LHOOQtius_ov_Borg]

A good visualization allows a good set of data to more readily be converted into the knowledge the user is seeking by analyzing the data. It adds information to the data through the visual arrangement, allowing humans - who are better at visual pattern recognition than numerical or textual pattern recognition - to quickly mine out the information they need.

Wile the Antarcti.ca idea is cool, and the design is visually attractive (at least, to me), I find that the presentation of the information could use some additional work. The sites that are listed are plotted as points on a map of "Antarctica" representing the Net - but the distance between points and their relative positioning does not seem to carry any particular meaning (at least, the sites I linked to from this site did not seem to have easily discernable traits which would make the distances meaningful)

The site is certainly attractive, but I think the integration of a system of text similarity measurement and positioning of the sites based on these metrics would greatly enhance the utility of the site (or, if this is already being done, it is not being done very well and needs improvement)

Also, the site will not be able to grow very large without a better notion of clustering based on some actually meaningful site comparison metrics (text similarity, google-like popularity measurements, etc.), since in order to have millions of sites on the map the idea they have implemented of moving around the map will need to be combined with some notion of places on the map having greater semantic meaning and also to reduce clutter, a more dynamic system of display which allows for better "zoom in / out" type functionality to "drill-down" into areas of interest based on representative, canonical sites that would appear on the high-level map.

All the beginnings are there, and it is a good start, I hope this goes somewhere...

By the way...
Anyone who thinks this field is cool, check out (if you haven't already) the books by Edward Tufte (I've put a reference at the bottom). And if you think you'd like to make some competing visualizations, check out this really great start-up data relationship viz tool company which my company is considering partnering with: www.thinkmap.com
----------------
Tufte's Books:
http://www1.fatbrain.com/asp/Search/SearchResult s.asp?from=bookinfo&Au=Tufte%2C+Edward+R%2 E&Ti=&Su=&Pu=&RegAction=t&SearchFunction=reg&qorde r=title