Ci Values as Keys
One problem raised early in the implementation:
-
the Hilbert numbers produced by Ci
could be expanded to arbitrary precision in order
to distinguish between
Ci(p1 )
and
Ci(p2 )
To use such values as dbm keys, we
need them as fixed precision, so we limit
expansion (to 4 levels).
Problems:
-
gives keys that are 96 bytes long
(producing very large Index files)
-
different vj can be mapped to same point
(size of candidate sets is artificially inflated)
Solution:
-
use a "prefix" index structure (something like a trie)
|