Numberless, tensed language for action oriented tasks
Action oriented languages are number intensive. Graphic's languages are centered around where to draw something rather than what to draw. The where involves a tedious numeric description of vertices. Robotic's languages are also dominated by a where description, but now the where specifies a robot motion. The result is an array of numbers that obscures the meaning of the program to its reader. This paper shows how a number of linguistic devices can be used to eradicate the plethora of numbers from action oriented descriptions. Functions or verbs can be tensed (e.g., past tense) to modify their meaning without duplicating the root function. The result is an English-like description to a control structure. Arguments or nouns can be modified in name, like the use of a GENSYM function in Lisp which generates a unique variable name from a character string, and in number (e.g, singular vs. plural). The result is an English-like description of bound and quantified variables. The remaining quantitative description of action tasks can be relegated to a database whose management system is specialized for number management. The resulting language is a formal variant of a natural language with a Lisp-like syntax (i.e., lists with functions in the first position). The programs approach the readability of a natural language without the cost of ambiguity that is inherent in natural descriptions. Finally, the programs can be easily pretty printed in English so that they can be read by non-programmers.
- Research Organization:
- Carnegie-Mellon Univ., Pittsburgh, PA (USA). Robotics Inst.
- OSTI ID:
- 7091760
- Report Number(s):
- AD-A-126400/1
- Country of Publication:
- United States
- Language:
- English
Similar Records
Selecting tense, aspect, and connecting words in language generation
Assessing the Linguistic Productivity of Unsupervised Deep Neural Networks