Given recent discussion let's reconsider the
proposal we've put up for word masking. Should we
withdraw it? After all, it's purpose was to
support cql searches, and if we're back to
defining word indexes, we don't need it, as long
as we have a way to map to Z39.50 attributes --
which we do for attribute architecture but not for
bib-1.
So I suggest the following:
1. withdraw the word masking proposal.
2. propose three new structure attributes for
bib-1:
left-truncation on word boundary
right-truncation on word boundary
adjacentWords
3. If these attributes aren't accepted then we
just map to attribute-architecture-friendly
combinations.
--Ray
