[Sigia-l] Information Visualization

[Listera]

"Karl Fast" wrote:

> However, what I've been reading (apparently misreading) from your
> posts is that size matters, infoviz won't succeed unless it deals
> with the size issue, and yet it can't do that because it's basically
> impossible or requires some unknown techno-magic.

For infoviz to get traction it needs acceptance and actual use by a sizeable
audience. A few users at small commercial and academic sites will keep it in
perpetual obscurity.

The entities most likely to fund infoviz development and drive its greater
exposure are in fact the owners of large datasets: they not only have the
money but also the acute need to make better sense of their data mountain.
However, they won't touch infoviz until and unless they are convinced that
it can scale, without which its usefulness to them is severely limited.

The colorful infoviz demos I've seen (and actually been involved in more
than a decade ago) are not designed to scale. It'd be suicidal to try to
sell commercial/shrink-wrap apps that come with a "Does not scale well"
sticker. :-)

Imagine that the web browser came with a warning: can only display a total
of 100Kb of data, has a request-response latency of 20 secs max, can only
open one window at a time, can only connect to servers within six hops, etc.

Why am I harping on scalability? It's both a conceptual as well as a
technical problem. It seems to me that many of those who design infoviz
contraptions haven't taken the time to think through ramifications of the
latter. And I have a suspicion that they don't have serious intentions to
sell their tools to large groups of users or those with large datasets
anyway. 

Of course there's no requirement that all infoviz tools should scale to
appeal to larger constituents. But (I assume) we are talking about the
future of infoviz *as an industry*. Infoviz, in general, is not going to
capture the imagination of the masses by staying small-scale and attacking
only non-scalable problems, which doesn't pay the bills.

If Google didn't scale well, we wouldn't have a search industry that can now
cash in its ubiquity/scalability for advertising dollars in the hundreds of
millions of dollars. On the other hand, Akamai doesn't scale well
*cost-effectively* and we don't have a comparably ubiquitous streaming
industry.

So for infoviz to leave its perpetual "rehearsal" mode and enter the
commercial arena, it needs to be (and be seen as) scalable.

However, as I said, scalability is only *one* of the problems. There are
cognitive/usability/usefulness issues beyond merely technical ones.

Ziya
Nullius in Verba