On Mon, Aug 24, 2009 at 6:54 PM, David Gerard<dgerard(a)gmail.com> wrote:
2009/8/24 geni <geniice(a)gmail.com>om>:
In practice the end point of Natural language
processing and large
scale digitalisation is likely to be made to request computer
generated custom articles. Wikipedia with it's surprisingly structured
entries is likely to be used as a significant stepping stone in this
direction.
Yes. Rather a lot of conversations I have with random geeks are along
these lines.
Basically, infoboxes = machine-readable data = good. So the way to go
there would be to make template plumbing give data more amenable to
chewing on. (Standardised names for template parameters, e.g. name,
photo.)
Surely the real information is in the semantic linkings? The concept
of the wikilink is still in its infancy, I think. If linking can be
optimised (by editors), it will be fascinating to see what
relationships that throws up.
Carcharoth