Illustration

 

This section contains a step-by-step walkthrough from a DATR theory to a trace file. All files discussecd here can be found in the directory zdatr/ example/. The example DATR theory is engcomp.dtr and was developed in a class held at the University of Bielefeld by Dafydd Gibbon

1. Given a DATR theory, zdatrtok will preprocess all equations, variables, atoms and nodes. To start the program, at least the input file name is needed:

     zdatrtok -i engcomp.dtr

   Because no token file name was given to zdatrtok, the token file output is written to the file with the automatically generated name engcomp.dtr.tok.

   Neither the DATR theory nor the token file will be listed in this document: the theory file is too long (it should be viewed using an editor) and the token file is quiet cryptic and not of general interest.

2. When zdatrinf is used interactively with manual, it is called with a token file only, e.g.:

     zdatrinf -t engcomp.dtr.tok

3. When zdatrinf is not used interactively with manual input, a query list file or a declaration file is needed. This example will use both simultanously. The query list file engcomp.qry (see 6.1) is just a list of queries. By using the information provided in the declaration file engcomp.dec (see 6.2) zdatrinf is able to construct queries and furthermore make conditional queries. In order to observe the results in detail, zdatrinf will be invoked with a trace file name and at verbosity level 1:

     zdatrinf -i example.dtr.tok -d example.dec -q example.qry -t example.out -v 1

   By invoking zdatrtok the two files zdatr.log (see 6.3 and zdatr.err have been created. zdatr.err should be empty for this example, because no error should have appeared.

 
• Query file engcomp.qry
• Declaration file engcomp.dec
• Log file zdatr.log