To view or change various processing options. 

The correct setting of processing options are important. Incorrect settings might cause the KB not to behave as desired. 

• Check update/delete integrity violations:
  • If checked: Record updates/deletes which could result in integrity errors are allowed only if relevant records are deleted or suspended. Note in this regard that updates to a record generally is processed as delete old record followed by insert updated record. The first step in this transaction can typically cause referential integrity errors.
  • If not checked: Update/delete of record will take place unchecked.
• Check for duplicate sentences:
  • If checked: Pro/3 checks if an identical record exists before a new record is inserted. The insert is ignored if the record exists already.
  • If not checked: Insert will take place unchecked. The purpose of skipping this check is to improve performance during mass updates. The performance improvement can be very significant. However, you should be certain that the sentence rules are defined such that duplicates are avoided.
• Check for other errors
• Update indexes:
  • If checked: All indexes are updated each time the KB is changed.
  • If not checked: No indexes are updated. Note in this regard that there is a command (in the KB menu) to re-generate all indexes. The purpose of turning index update off is to increase performance in connection with mass updates of the KB.
• Update dependency graphs:
  • If checked: All dependency graphs are updated as required each time the KB is changed.
  • If not checked: No dependency graphs are updated. Note in this regard that there is a command (in the KB menu) to re-generate all dependency graphs. The purpose of turning dependency graph update off is to increase performance in connection with mass updates of the KB.

Sentence rules

• Save new rules as inactive or Save new rules as active determines the record status of newly inserted sentence rules.
• Populate rule editor identifier boxes with true identifiers (else with all) determines if only true application identifiers will be used to populate the KE IDF (and STD IDF) listboxes in the rule editor. All identifiers will mean that entity type names, predicate type names etc are included in the listbox. These are normally not used as identifiers in rules.
• Populate rule editor's DE listbox with Pro/3 standard DEs (else only with DEs defined by the application).

• Store inexact rule reasoning trees:
  • If checked: The returned value for executions of inexact rules called by sentence rules, are stored as sentences in the KB. These sentences can be used to study the reasoning carried out. The sentences are also the basis for the drawing of certainty rules reasoning trees. See inexact rules for details.
  • If not checked: No inexact rule reasoning sentences are stored in the KB.
• Store inexact rule reasoning details:
  • If checked: The returned value for all executions of inexact rules are stored as sentences in the KB. These sentences can be used for detailed analysis of the behavior of the inexact rules.See inexact rules for details.
  • If not checked: No inexact rule reasoning sentences are stored in the KB.
• Date formats - see date and time domains.

Time zone

• Hours and minutes relative to GMT.

Unicode

• File format for knowledge source (input) files (Pro/3 always output in UNICODE).
• Show Usage Guide (else not).
• Automatic KB Backup (else no automatic backup).
• Re-scan threshold - NL-scanning is non-deterministic and Pro/3 retries until a valid interpretation of the syntax is found. The threshold defines the number of retries which when reached will prompt the KE to confirm continued scanning or else abort the syntax analysis.
• Save preference settings at exit:
  • If checked: All prefence settings and options will be saved to file each time you exit PRO3.EXE. The file has type .INI and the same name and directory as the KB.
  • If not checked: The options and preferences are not saved to file upon exit from PRO3.EXE. Note in this regard that there is a command (in the Preferences menu) to save the current preferences and options to file.

TRACE PREFERENCES

Trace data is written to a trace file with type .3tc and the same directory and name as the KB. Trace is used while testing and debugging your Pro/3 model. There are two main classes of trace: translation trace and inference engine trace. Translation trace covers the process of translating knowledge input to IF. Inference engine trace covers the process of processing queries (sentence derivations).