Command and Natural Language

1
Command and Natural Language

• MSIT 159 User Interface Design and Development

2
Design Goals for Any Language

• Precision
• Compactness
• Ease in writing and reading
• Speed of learning
• Simplicity to reduce errors
• Ease of retention over time

3
Higher Level Language Design Goals

• Close correspondence between reality and language
  notation
• Convenience in carrying out manipulations
  relevant to user tasks
• Compatibility with existing notations
• Flexibility to accommodate both novice and expert
  users
• Expressiveness to encourage creativity
• Visual appeal

4
Constraints on Language

• Capacity for human beings to record notation
• Match between the recording and the display media
  (for example, clay tablets, paper, printing
  presses)
• Convenience in speaking (vocalizing)
• Successful languages evolve to serve goals
  within the constraints.

5
Functionality to Support User's Tasks

• Users do wide range of work, e.g.
• text editing
• electronic mail
• financial management
• airline or hotel reservations
• inventory
• manufacturing process control
• gaming

6
Designer should

• determine functionality of the system by studying
  users' task domain
• create a list of task actions and objects
• abstract this list into a set of interface
  actions and objects
• represent low-level interface syntax
• create a table of user communities and tasks,
  with expected use frequency
• determine hierarchy of importance of user
  communities (i.e. prime users)
• evaluate destructive actions (e.g. deleting
  objects) to ensure reversibility
• identify error conditions and prepare error
  messages
• allow shortcuts for expert users, such as macros
  and customizing system parameters

7
GOMS Model for Command Language Interfaces

• Basic goal - speed and flexibility
• Basic method
• Step1. Think of and enter command verb
• Step2. Think of and enter next argument
• Step3. if more arguments then go to Step2
• Step4. if command is incorrect then correct the
  command
• Step5. Signal computer to process the command.
• Step6. go to Step1

8
What makes a command language easy to learn and
use?

• Easy Command Synthesis
• User can think up command by analogy from
  previously learned commands
• Commands conform to simple rules rather than lots
  of unique special cases

9
Command Language Organization Strategies

• Simple command set
• Commands plus arguments
• Commands plus options and arguments
• Hierarchical command structure (command action,
  object argument, destination)

10
Benefits of Structures

• Ease of learning, use, and recall
• Easy to construct guidelines documents
• Permits discussion of impact of allowing
  exceptions

11
Command Language Abbreviation Strategies

• Simple truncation
• Drop vowel and use simple truncation
• First and last letters
• Standard abbreviations from other contexts
• First letter of each work or phrase
• Phonics (e.g. XQT for execution)
• Cross product languages verbs x objects

12
Command Language Guidelines

• Make command terms easy to remember
• Provide easy command synthesis method and
  abbreviation strategy
• Provide simple, consistent command structure
• Commands should be right grain size
• Parsimony (no more commands than really needed)
• Studies show a few commands are used a lot by
  most users
• Provide for command reuse (replay, re-entry,
  macros)
• Avoid unnecessary distinctions among commands

13
Natural Language Processing

• Eliza (Simon Laven Chatterbot page) Developed by
  Wizenbaum simulated Rogerian therapy, there is no
  real understanding, only a lot of clever tricks.
• Pattern matching (primary and secondary)
• mechanical rewriting of statements into questions
  
• key word vocabulary (e.g. mother, hate, love)
  used to focus attention
• remembered concept (when deadends are reached)
• Early investigators thought that grammar (or
  syntax) was all you needed to make an automatic
  langauge translation system work. This was found
  to be wrong you also need to consider semantics.
• time flies like an arrow fruit flies like a
  banana

14
Context Free Grammars

• Context free grammar can be represented and
  parsed by finite state machines.
• These may be extended to form recursive
  transition networks (RTN), which might be thought
  of as FSM with networks labeling the arcs. This
  allows the nodes to call subroutines (context
  sensitive grammars).
• Augmented transition networks (ATN) add registers
  to RTN to store partially developed parse trees,
  allow conditional execution of arcs, and attach
  actions capable of modifying the data structure
  returned. This means that ATN's are Turing
  machines.

15
Conceptual Dependency

• Roger Shank's approach to representing deep
  meaning.
• I gave the man a book
• ATRANS Abstract transfer (give)
• PTRANS Physical transfer (go)
• MTRANS Mental transfer (tell)
• P Past tense PP object or picture producer
• 0 Object case relation AA Action modifier
  (aider)
• R Recipients Act Action
• PA Picture modifier (aider)

16
Conceptual Dependency (contd)

• How can conceptual dependency facilitate
  reasoning?
• Fewer inference rules.
• Many inferences are contained in the
  representation
• Initial representation of the sentence will have
  holes. Plugging holes serves as focusing a
  subject for future sentences.
• Argument Against
• Long time to decompose knowledge into primitive
  actions.
• Conceptual dependency is good for representing,
  this may not be good for all kinds of knowledge.

17
How does Natural Language Processing (NLP) fit
into user interface design work?

• In natural language interface (NLI) queries are
  open ended prompts like "what do you want to do?"
  Which gives very little support to guide user
  actions
• Might be used in museum type applications to
  allow natural language queries (NLQ) against
  relational databases. NLQ is parsed and
  translated to standard SQL.
• Search engine searches against text database
  (e.g. find cases where tenants sue landlords. May
  allow modified SQL type query to search against
  simple domain model.
• Natural language text generation. Reports might
  be written automatically from lab data or output
  might make use of computer generated speech to
  help the visually impaired.
• Text-based adventure games. Restricted domains,
  make parsing a little easier than continuous
  speech recognition. Parser is a little more
  sophisticated than Eliza.

18
Turing Test

• The classic test of machine intelligence is to
  have a person communicate electronically with two
  entities (one a machine and the other a person).
• If the inquisitor is unable to determine who is
  the machine and who is the person "true
  artificial intelligence" has been exhibited.
• Each year a restricted version Turing
  competition is held at the Boston Computer
  Museum.
• Some progress is evident, but more work needs to
  be done.