|
|
| Historic Environment Resources Network | http://www.britarch.ac.uk/HEIRNET/ |
| Home | Technical Advice | | |
Z39.50 FAQs
What is Z39.50? Back to the topIn simple terms, Z39.50 is no more than a network protocol: it is a set of rules that controls how two or more computers talk to each other over the Internet. Z39.50 was developed to support networked access to distributed databases, and it allows for searching and retrieval of information from databases at different geographical locations through one user interface via the Internet. Who uses Z39.50? Back to the topZ39.50 has been used most often by libraries, as a means of retrieving bibliographic records from catalogues maintained by different libraries simultaneously. It has also been used in the UK in the Higher Education sector and in museums, to provide combined access to distributed collections of electronic data. HEIRPORT uses Z39.50 to search different databases which each contain information about the historic environment. The development of HEIRPORT is significant because it provides users with access to resources that are held by organisations who are located across the UK. Using a standard web browser, users can iniate searches of the target databases (which are held in different database management systems on various hardward platforms) - the results are passed back to users' browsers and presented to them in HTML. The good news for users is that in implementations of Z39.50, the technology is effectively hidden, and as far as they are concerned they're simply searching one great big database. How does Z39.50 work? Back to the topFor the more technically minded, Z39.50 follows what is known as a client/server model, where one computer (the client) submits a request to another computer (the server) which services the request and returns some kind of answer. In terms of Z39.50, the client is called the 'Origin' while the server is called the 'Target'. Queries are sent over the network by the origin to the targets in z39.50-speak. The target does the work of translating the query into proprietary-database-speak (Oracle, SQL, etc) using a 'Profile', which specifies how data and queries are to be handled in the particular domain. The query is then executed by the database and any results are returned to the target, which uses the profile to translate the results back into z39.50-speak and then sends them to the origin. The origin then does the work of translating results from each of its targets into a suitable format for presentation to users. Paul Miller's article, 'Z39.50 for all' provides a good introduction to Z39.50: What about my database? Back to the topThe advantage of Z39.50 is that each target is independent. Z39.50 does not specify what sort of database management system or which type of hardware your system should use. Nor does it specify the structure of your database - or its content. The way that Z39.50 works (see above) means that the staff who maintain your database can develop and enhance the data structures that are most appropriate to your working context. In theory any database can act as a target in a Z39.50 portal as long as they are mapped to a standard profile. In practice searches work better if the targets hold similar information, for example HEIRPORT targets historic environment databases which contain information such as site name, location, monument type and period somewhere in their data structure. The exact data fields and searches that are required to return this information can be mapped to standard profiles used by HEIRPORT. What sort of system do I need to be running to act as a target for HEIRPORT? Back to the topTheoretically any database management system can act as a target for HEIRPORT. At the time of writing targets have only been implemented for Oracle(registered trademark symbol), System Simulation Limited's Index+ and MySql database servers. Packages such as Microsoft Access (registered trademark symbol) may be implemented if only moderate usage is anticipated. HEIRPORT needs to be able to communicate with your system to run a session. In real terms this means that your system needs to be accessible to the global network 24 hours a day. How do I turn my server into a target for HEIRPORT? Back to the topTwo things are essential. Firstly, you need to install some software to allow your server to communicate across the network with HEIRPORT. Secondly, the data within the target database has to be mapped on to profiles set out in HEIRPORT's target specification - mapping describes your data in a format that the HEIRPORT origin can manage. What software will I need in order to become a target? Back to the topOne option is to develop your system to conform to the requirements of the HEIRPORT Z39.50 target specification. System Simulation Ltd took this route in setting up the Portable Antiquities Scheme database as a target; other users of their Index+ database may be able to take advantage of this development. Alternatively, you can take advantage of development work commissioned by the Archaeology Data Service. Zava, an Application Programming Interface (API) was developed by the University of Kent Computer Laboratory. Zava is being made available by the Archaeology Data Service as open source software to interested parties. If you use Zava, the following will need to be installed:
Each of these elements needs to be installed on the host system and tested to make sure that they work together. The RDFs are documents which hold the profiles, XML tag libraries and other metadata used by the target. See the Z39.50 Glossary from UKOLN online: http://www.ukoln.ac.uk/dlis/z3950/glossary/ What is the HEIRPORT Z39.50 target specification? Back to the topThe HEIRPORT z-target specification defines how communications will take place between your target and HEIRPORT and specifies conformance with two widely used standards:
HEIRPORT specifies conformance with both profiles as, taken together they offer the best searches for historic environment data. The use of XML is advantageous because it allows results to be displayed on a wide range of devices (web browser, WAP phone, TV Internet, etc). The HEIRPORT specification can be viewed at
http://ads.ahds.ac.uk/project/target_spec.html Can I customise my target to protect sensitive site data? Back to the topWhen you map the data in your database against the profiles specified by HEIRPORT, you decide which data can be used in searches and which is listed in the results. For example, you can specify that grid references may be used in searches but prevent the whole reference from being displayed in the results. Can I allow my target to be searched by more than one portal? Back to the topOnce your target is set up it can be searched by one or more origins. The Zava API developed by the Archaeology Data Service allows you to specify which portals can search your target. You might decide to restrict access to your target to HEIRPORT and other specified portals to limit the number of concurrent users of your database. Where can I go to get more advice? Back to the topThe Archaeology Data Service technical team can offer advice if you are considering setting up your database as a target for HEIRPORT. For more information e-mail afa2@york.ac.uk or email to heirnet@ads.ahds.ac.uk . ReferencesPaul Miller, 1999, 'Z39.50 for all', Ariadne issue 21, online: http://www.ariadne.ac.uk/issue21/z3950/intro.html Z39.50 Maintenance Agency online: http://lcweb.loc.gov/z3950/agency/ 'HEIRPORT: About the portal' information page online: http://ads.ahds.ac.uk/heirport/about.html The HEIRPORT specification can be viewed at: http://ads.ahds.ac.uk/project/target_spec.html |
|