The problem with the previous example is that you need to know the
structure of the documents in order to find them. For example,
when we wanted to find the record for the taxon
Sauroposeidon,
we had to formulate a complex XPath
/Zthes/termName
which embodies the knowledge that taxon names are specified in a
<termName> element inside the top-level
<Zthes> element.
This is bad not just because it requires a lot of typing, but more
significantly because it ties searching semantics to the physical
structure of the searched records. You can't use the same search
specification to search two databases if their internal
representations are different. Consider an alternative taxonomy
database in which the records have taxon names specified
inside a <name> element nested within a
<identification> element
inside a top-level <taxon> element: then
you'd need to search for them using
1=/taxon/identification/name
How, then, can we build broadcasting Information Retrieval
applications that look for records in many different databases?
The Z39.50 protocol offers a powerful and general solution to this:
abstract ``access points''. In the Z39.50 model, an access point
is simply a point at which searches can be directed. Nothing is
said about implementation: in a given database, an access point
might be implemented as an index, a path into physical records, an
algorithm for interrogating relational tables or whatever works.
The key point is that the semantics of an access point are fixed
and well defined.
For convenience, access points are gathered into attribute
sets. For example, the BIB-1 attribute set is supposed to
contain bibliographic access points such as author, title, subject
and ISBN; the GEO attribute set contains access points pertaining
to geospatial information (bounding coordinates, stratum, latitude
resolution, etc.); the CIMI
attribute set contains access points to do with museum collections
(provenance, inscriptions, etc.)
In practice, the BIB-1 attribute set has tended to be a dumping
ground for all sorts of access points, so that, for example, it
includes some geospatial access points as well as strictly
bibliographic ones. Nevertheless, the key point is that this model
allows a layer of abstraction over the physical representation of
records in databases.
In the BIB-1 attribute set, a taxon name is probably best
interpreted as a title - that is, a phrase that identifies the item
in question. BIB-1 represents title searches by
access point 4. (See
The BIB-1 Attribute Set Semantics)
So we need to configure our dinosaur database so that searches for
BIB-1 access point 4 look in the
<termName> element,
inside the top-level
<Zthes> element.
This is a two-step process. First, we need to tell Zebra that we
want to support the BIB-1 attribute set. Then we need to tell it
which elements of its record pertain to access point 4.
We need to create an Abstract Syntax
file named after the document element of the records we're
working with, plus a .abs suffix - in this case,
Zthes.abs - as follows:
