Mike Taylor wrote:
> So I herely propose a BIB-1 truncation attribute, 105, defined as
> "CQL-like wildcard matching".
Some considerations. A couple months ago there was strong sentiment among
us to hold off on such a proposal. Now that we have 1.0 it's reasonable to
revisit the question.
What's the best approach? Let's say we put up a proposal now with the cql
wildcard matching (though I'd call it 106, not 105). It's sure to be kicked
around, and we could use the ZIG input to help guide decisions for changes
in 1.1. At some point though we may have to say that the ZIG consensus in
this case isn't useful for CQL, at which point we might drop the proposal.
Or we could take a "take-it-or-leave-it" approach defining a type 106 how we
want it for cql, without incorporating ZIG consensus. That wouldn't be my
favorite approach but we can do it if it's sufficiently important for
SRW/Z39.50 implementors to have the cql masking available as a Z39.50
attribute.
Another approach is to hold off awhile until we have a better idea about
what if any changes might be in 1.1.
--Ray
|