Re: [LogiLogi-list] Introducing myself as part of the RU team
Status: Beta
Brought to you by:
wybow
|
From: Wybo W. <wy...@lo...> - 2007-04-15 22:38:05
|
A small note...: > Also, I plan to derive a mathematical specification of the notion of > concepts (etc) in LogiLogi, in order to get a better understanding of > the formal properties. My hypotheses is that the distinction between > primary and secondary tags is superfluous for the underlying model. > But as I said, this is something I need to investigate. If you imagine the primary and secondary versions of tags, as being two sets, then it immediately becomes clear that the distinction is not superfluous. (x in 'b/a' is added to b2, a2, and a1, y in 'a/c' to a2, c2, and c1, if you request 'a', all in a1 is returned, so only x) Sets are un-ordered, and by making a distinction between primary and secondary tags, we can have one layer in a simple way (the initial plan was much more complex - roughly creating a new concept for each new combination of words in an url (& it also involved suffix-trees) - but for pragmatic and esthetic reasons I dropped that and went for taggings of two types). Wybo |