Although the Gale group makes many of its databases available by Z39.50 (Business and Company Resource Center, Health Reference Center Academic, etc.), four of its databases that are required for the LOT are not available in this way.
These databases are available by means of a proprietary protocol that works by sending rigorously specified HTTP GET requests, receiving XML documents, and parsing them to discover the result of the request. The protocol is called HITS (HTML InfoTrac Server, a singularly inappropriate name since HTML is not involved at all.) Detailed documentation of the HITS protocol may be obtained from Gale technical support.
Support for this protocol in the LOT was added in two stages.
First, we created an open source Perl module, HITS.pm, for accessing databases via HITS using a simple object-oriented API.
Then we built a gateway, zHITSgate, by gluing the HITS API onto the back end of the generic Z39.50 server provided by the open source Net::Z3950::SimpleServer module (available at http://www.indexdata.dk/simpleserver/)
The great bulk of the code for the Z39.50-to-HITS gateway is contained in the HITS and SimpleServer modules (and the modules that they, in turn, use). This means that the zHITSgate script itself is quite manageable - it's about 340 lines of Perl, including comments. Apart from translating Z39.50 queries to HITS queries, and enhancing HITS XML records with Dublin Core elements that the LOT software knows about, its main job is to provide configuration information specifying what databases to support, what access points they know about, what DC elements are included in their records, etc.
This means that, in order to add support for additional HITS databases (or change the details for those already supported), the zHITSgate script itself is the only file that needs to be edited. The underlying HITS and SimpleServer modules do not need to be changed.
When connecting to zHITSgate, a Z39.50 client should provide authentication - a username and password - to be passed through to the HITS server. If this is omitted, the Z39.50 Init request will be refused. It is a matter of policy what authentication information is used: a TexShare-global username/password pair might be used by all LOT access, or a library-specific username/password pair depending on which local library the user specifies. The gateway doesn't care - it just passes through whatever authentication information it is given.
The current version of zHITSgate supports the databases described in the following sections:
Vendor: Gale
Z39.50 database name: BGMI
This is not a bibliographic database, but a database of people. Accordingly, a title-search actually searches for people's names, since the names are ``titles'' of the records.
Table 5-2. BGMI: access points supported
| Access Point | Meaning | Notes |
|---|---|---|
| 1003 | author | Searches last name only |
| 1016 | Keyword | Full-text search |
| 1032 | Doc-id | As returned as Dublin Core identifier |
Table 5-3. BGMI: Dublin Core fields returned in brief/full records
| DC Field | Meaning | Brief | Full | Example |
|---|---|---|---|---|
| date | Birth/death dates | Y | Y | 1937- |
| type | Document type | Y | Index Record | |
| title | Name reversed | Y | Fruit, Melvyn Herschel | |
| title | Name | Y | Melvyn Herschel Fruit | |
| identifier | Record number | Y | DO1900123100 | |
| source | Source | Y | Biography and Genealogy Master Index. |
Vendor: Gale
Z39.50 database name: SRCG/SRC1
It appears experimentally that this database's full records contain two fewer Dublin Core elements than its brief records! (The full record also contains full text in most cases, or at least a long summary, but that information cannot be represented using Dublin Core fields.)
Table 5-4. SRCG: access points supported
| Access Point | Meaning | Notes |
|---|---|---|
| 1003 | author | |
| 21 | subject | |
| 1016 | keyword | Using this access point seems to be the same as not providing an explicit access point at all. |
| 1032 | Doc-id | As returned as Dublin Core identifier |
Table 5-5. SRCG: Dublin Core fields returned in brief/full records
| DC Field | Meaning | Brief | Full | Example |
|---|---|---|---|---|
| description | Annotation | Y | (Metro) | |
| rights | Online rights | Y | -- Text Available -- | |
| source | Journal Name | Y | Y | The Plain Dealer (Cleveland, OH) |
| date | Publication info. | Y | Y | Dec 3, 2003 pB3 |
| identifier | Record number | Y | Y | CJ94846107 |
| title | Title | Y | Proposed recreation cuts, fee hikes draw complaints. | |
| author | Author | Y | Y | Susan Morrison |
Vendor: Gale
Z39.50 database name: GTXS
Table 5-6. GTXS: access points supported
| Access Point | Meaning | Notes |
|---|---|---|
| 1016 | keyword | The Gale documentation says this works, but I can't get it to find any records. Better omit the access-point specification and use the default keyword search. |
| 1032 | Doc-id | As returned as Dublin Core identifier |
Table 5-7. GTXS: Dublin Core fields returned in brief/full records
| DC Field | Meaning | Brief | Full | Example |
|---|---|---|---|---|
| type | Document type | Y | Y | Table |
| identifier | Record number | Y | Y | CG2199000423 |
| source | Source citation | Y | <I>Texas Almanac, 2002-2003</I>. Dallas Morning News, 2001. | |
| title | Link text | Y | Texas Cattle Marketed, 1965-1999 by Size of Feedlot | |
| title | Document title | Y | (as above) |
Vendor: Gale
Z39.50 database name: WDRN
This database is actually a union catalogue of lots of different kinds of records, including things like lists of award winners as well as actual books. Accordingly, although there are a lot of fields that can appear, no single record seems to have anywhere near all of them.
Table 5-8. WDRN: access points supported
| Access Point | Meaning | Notes |
|---|---|---|
| 4 | title | Requires full title, not just a word from the title |
| 1003 | author | Requires full name, not just surname or forename |
| 21 | subject | May require full title |
| 1032 | Doc-id | As returned as Dublin Core identifier |
Table 5-9. WDRN: Dublin Core fields returned in brief/full records
| DC Field | Meaning | Brief | Full | Example |
|---|---|---|---|---|
| format | Book format | Y | ||
| subject | Subjects | Y | Poetry | |
| type | Document type | Y | Y | Awards List |
| contributor | Editor name | Y | ||
| description | Genre | Y | Inspirational - Westerns | |
| author | Name reversed | Y | Moore, John L. | |
| author | Name | Y | John L. Moore | |
| date | Publication date | Y | 1995 - 2001 | |
| identifier | Record number | Y | Y | M1300000070 |
| source | Source citation | Y | What Do I Read Next? Farmington Hills, Mich.: Gale Group. | |
| relation | Similar titles | Y | ||
| title | Link text | Y | National Poetry Series Competition | |
| title | Title | Y | (as above) |