Glossary

chaining

The ability to invoke multiple recognition elements in a row by speaking them as a single utterance, i.e. without pausing between them. In dragonfluid, chains may be of any length. Depending on the scenario, two neighboring chained elements may be comprised of free speech dictation and a command, in either order, or a pair of commands.

Only rules derived from ContinuingRule will pass off, or chain, to successive commands.

command action

The action executed when a rule is triggered by its command.

command

Spoken content within an utterance that is meant to trigger the execution of a command action. Often specified in the form of a spec. In contrast with dictation.

dictation
free speech dictation

Spoken content within an utterance that is meant to be captured as its textual representation, generally as a means to supply content to a command action, such as for printing to the screen. In contrast with a command.

dictation container

A type of value produced by a Dictation element, derived from DictationContainerBase, and specific to the speech recognition system in use. The elements provided by dragonfluid, such a SplitDictation can also return these containers upon request.

embedded command

A command within an utterance that does not occur at the beginning of the utterance.

extras

Broadly speaking, an extra is a part of a command that hears and results in a certain type of content.

An extra uses a named element, derived from ElementBase, and provides a value, such as text or a dictation container. Dragonfly documentation provides a list of elements.

It is often used here as a term for the dictionary of extras passed to the _process_recognition callback of a rule. You are generally expected to know how to access the various extras from this dictionary, and when documentation states that extras are passed to or returned from a function, the form implied is this dictionary. Note that this dictionary generally does not container the underlying element that generates a value, so extras are distinct from elements, with extras using elements and producing values under the same name.

intro
command intro

The initial part of a command’s spec, consisting only of static literal words, up to the first encountered extra reference in angle brackets. For further details see the intros concept section.

literal tag

A word spoken within an utterance to specific that what follows is free speech dictation even when it looks a command or literal tag. For further details see the literalization concept section.

registered command

A command that can be triggered from within the middle of an utterance. For further details see the registration concept section.

rule
macro

A triggerable event. The trigger is the command and the event triggered is the command action.

spec

A common dragonfly attribute that determines which spoken words will trigger a rule. It may be a fixed literal command spec, such as “show desktop”, or it may include references to extras in angle brackets, such as “delete left <characterCount> characters”.

utterance

The contiguous stream of spoken content captured by your speech recognition system starting from the moment your it determines you have begun speaking through until the moment it encounters enough silence to qualify as a pause given its configuration.