Major Changes to Bath Profile
in the
January 10, 2000 Revised Draft for Public Comment


1.  Removed the terminology in Level 0 conformance that explicitly stated this level as "easy-entry" level.

2. Remove qualifiers from the term "interoperability" except when trying to explicitly highlight the idea of "semantic interoperability".

3.  Revised the statement about what "support" means for Z-clients and Z-servers.

4. Clarifying the use of terms such as indexes, fields, data elements, etc. 

5.  Keyword searches containing multiple search terms

6. Changed semantics for cross-domain searches

7. Changed wording in cross-domain searches to remove "fields" and replace with "data elements".

8.  Character set support

9. Added 2 Relation attribute values to support Date searching  

10. Changed Structure Attribute in 5.A.1SCAN.1 Author--Exact Match

 



1.  Removed the terminology in Level 0 conformance that explicitly stated this level as "easy-entry" level.

Old:

Conformance Level 0 is considered an "easy-entry level" and is intended to encompass as many existing Z39.50 products as possible; conformance with Level 0 may require the reconfiguring of existing implementations, but it should not require product development 

New:

Conformance Level 0 defines requirements for a limited number of searches to improve semantic interoperability, and is intended to encompass as many existing Z39.50 products as possible; conformance with Level 0 may require the reconfiguring of existing implementations. 


2. Remove qualifiers from the term "interoperability" except when trying to explicitly highlight the idea of "semantic interoperability".


3.  Revised the statement about what "support" means for Z-clients and Z-servers.

Old:

"Support" in this context means:

New:

"Support" in this context means:


4. Clarifying the use of terms such as indexes, fields, data elements, etc. 

In an attempt to clarify this issue, there is the following paragraph inserted in Section 2, Purpose and Scope:

Terminology issues are ever-present in a document such as this.  An example of such an issue is with terms such as "access points," "indexes," "fields," and "data elements."  Often, the library community uses the phrase "search a particular field or fields," when at the system level, the search may be executed by matching the search term with entries in a system-generated index.  Access points can be considered searchable fields of a record as represented by the index created from data from those fields.  For Cross-Domain searching, the concept of "field" may be completely absent.  In defining searches for library catalogs, the description references fields and indexes.  In defining cross-domain searches, the description references data elements and indexes.

And specifically in Section 5.A., the wording has been changed as follows.

New: 

This profile does not specify data elements or indexes to be mapped to the required bib-1 use attributes.  It recognizes that indexing practices may vary based on local needs.  However, it assumes that in library catalogue implementations:

Old: 

This profile does not attempt to specify data elements to be mapped to the required bib-1 use attributes. It
recognizes that there may be variations based on local needs. However, it assumes that in library catalogue
implementations:


5.  Keyword searches containing multiple search terms.  Changed some wording suggestions by Gatenby to clarify client formulation of these keyword searches in Section 5A.

Old:

Keyword searches using more than a single-word search term requires Z-clients to formulate a
multi-term keyword query by using a Boolean operator.

New:

Where a keyword search contains multiple words, each word is a separate term with associated bib-1 attributes to form an operand within the query.  Searches with multiple operands  are combined a Boolean operator. 

This however doesn't address whether we are going to specify a specific Boolean operator (e.g., AND) for the searches.  See issues page.


6. Changed semantics for cross-domain searches.  To bring in line with current semantics from Dublin Core, 1.1.  For example

Old:

Uses: Searches for the year in which a resource is published. 

New:

Uses: Searches for the date (year) associated with an event in the life cycle of the resource.
Typically, this date will be associated with the creation or availability of the resource.


7. Changed wording in cross-domain searches to remove "fields" and replace with "data elements".

Example:

Old:

Uses: Searches for complete word(s) in fields that contain names of a people or entities responsible for
making the content of a resource.

New:

Uses: Searches for complete word in data elements that contain names of people or entities
responsible for a resource.


8.  Character set support:  Added the following to Section 4.3.3., Retrieval: Record Syntaxes.

Interoperability requires use of  standard character sets.  Unless specified otherwise, Z-clients and Z-servers must use the character set defined by a particular record syntax (e.g., for MARC21, see MARC 21 Specifications for Record Structure, Character Sets, and Exchange Media).

Do we want to add additional information about the character sets?  And should we state explicitly someplace that origins and targets are not precluded from using other characters, but the profile prescribes the ones that all implementations will support?


9. Added 2 Relation attribute values to support date searching

Relation (2):   1, 2, 3, 4, 5:   less than, less than or equal, equal, greater than or equal, greater than


10. Changed Structure Attribute in 5.A.1SCAN.1 Author--Exact Match

To maintain parallelism with the searches defined in Level 1, the Structure Attribute for this Scan should be value 1, Phrase, and not 101, Normalized Name.  See open issue regarding the requirement for Structure Attribute 101 in any of the searches.


 

 

[end]