Monday, October 11, 2004
Manhattan Project to Integrate
Human-centric Information Production
This is a historical document as Tom Adi’s first communication to the
BCNGroup on the formal structure of the Readware framework
Second, more extensive communication is at [65]
As I review the proposal, here is an overall semi-formal structure of the
Readware Framework:
1) Everything is based on relations of the type <a,r,b>.
Concept = Arabic letter triplet = <LetterA, relationLetterB,
LetterC> (there are special cases).
Knowledgetypes (KT) is either topic, issue, probe, category or theme:
KT = <conceptA, InContextOfSizeM, conceptB>
OR recursively
KT = <KT1, InContextOfSizeM , KT2)>
2) The abstract space is
A = {(element, domain, order) x (unitary open, unitary closed, dual open,
dual closed) x (object, relation)}
A concept is a member of A^3 (A x A x A).
3) Some 2000 Arabic concepts are used as a good cover of A^3 (not all of
it makes sense). But we can add more triliteral word roots from Arabic or
any other language. It's just a lot of work while the stem set we use is
stable, well-interpreted and sufficient.