Mailing List Archive

Axel Boldt wrote:
> When all is said and done, I would like a picture like
> http://ontology.teknowledge.com:8080/rsigma/sumo-1.36classes.pdf
> and when I click on a category, I'm sent to a category page which lists
> the immediate superclasses and subclasses (editable), gives an editable
> description of the category, gives a link to the recent changes in that
> category (including all changes in its subcategories), a link that
> allows searching in the articles of that category and its
> subcategories, and a link to a complete alphabetical list of all
> Wikipedia articles belonging to that category and its subcategories.

Every wiki page is potentially a category and wiki links could be
sorted into "is-a" (from instance pages to category pages), or
"superclass" or "subclass" links (between category pages). If you
introduce typing in the wiki link syntax, a program could find all the
superclass-subclass links and draw the diagram or use the link tree
as the basis for your recent changes listing. For example, the link
from [[Dolphin]] to [[Mammal]] would be a superclass link, which could
be written: "A '''dolphin''' is a [[superclass:mammal]]".

Now, I don't suggest we introduce this sort of link typing, because I
think that would only lead to a bigger problem rather than providing a
solution. What about the link between Edison and the phonograph,
should there be an "inventor_of" (and "invented_by") link type? Did
Guglielmo Marconi really invent radio, or do we need plain English to
describe the vagueness that most consider him to be the "father of
radio"? My personal opinion is that any attempt at a hard encoding of
knowledge is in fact inferior to plain language and untyped links.

(The discussion that led to the current, deliberately vague wording of
[[Guglielmo Marconi]] took place before June 2001, and is summarized
on [[talk:Guglielmo Marconi]].)


--
Lars Aronsson (lars@aronsson.se)
Aronsson Datateknik - http://aronsson.se/