Dr. Murai Éva - Gubányi András szerk.: Parasitologia Hungarica 27. (Budapest, 1994)

unique" feature in the case of parasites recovered from the same host. It means two databases are linked, which have different characteristics considering the permission of multiplication of records with the same field content in a file. OVERVIEW OF PARDATA A programme called ParData program has been written in a professional database management language, named Clarion (version 3.0). The Clarion language is pragmatic - a programme language designed with convenience in mind. Clarion uses a file organization that has become common among desktop database manager products: data records are contained in a file along with a header that describes the file and record layout. Each record key is contained in a separate file. Clarion files are robust. Between accesses, they are logically closed making them invulnerable to power failures and reboots. Clarion showed the best results among different database management languages after comparing computing speed, the executable programme size and the transaction processing of them (Table 1). Clarion "Sieve of Eratosthenes" programme (an algorithm for finding prime numbers) also run twice as fast the same programme written in Borland's Turbo C++. Table 1 Comparison of different computer languages Sieve of Eratosthenes computing speed Hello word.exe size Transaction / second Clarion 6.2 sec 10 K 25.5 Clipper 10 min 34 sec 141 K 8.0 Foxpro 16 min 10 sec 656 K 6.4 STRUCTURE OF PARDATA The main concept was to develop a code system which can reduce the size of the database files. Thus the records of the main files do not contain string variables and the speed of information processing will be increased. All the string fields are deposited in auxiliary files and recalled automatically from them. Taking taxonomical points into consideration besides the widespread criteria, one of the most important requirements for a zoological database is flexibility to nomenclatural changes. A code system seems to be suitable for defining such a hierarchical architecture flexible to nomenclatural changes. There are two types of relationships that are difined in the ParData database: One-to-Many and Many-to-Many relationships.