TEACH SEM1A3                                 Aaron Sloman September 1995

                            THE SEM1A3 COURSE
              Artificial Intelligence Programming in POP-11
                          (Year 1, Semester 1)

This is the practical component of the first semester of the
AI degree course. The course is assessed by a mini-project to be done at
the end of the semester. It is followed by a course in Prolog
programming (SEM1A4).

CONTENTS

 -- OBJECTIVES OF THIS COURSE
 -- TIME ALLOCATION 6.5 hours per week per module
 -- SPECIFIC SYLLABUS
 -- 1. LEARNING TO USE THE POPLOG DEVELOPMENT ENVIRONMENT.
 -- 2. INTRODUCTION TO UNIX.
 -- 3. PROGRAMMING AND POP-11
 -- 4. THE IDEA OF SOFTWARE ENGINEERING DISCIPLINE
 -- 5. AI CONCEPTS AND TECHNIQUES
 -- 6. MINI PROJECTS
 -- 7. PROVISIONAL COURSE TIMETABLE
 -- -- Week 1
 -- -- Week 2
 -- -- Week 3
 -- -- Week 4
 -- -- Week 5
 -- -- Week 6
 -- -- Week 7
 -- -- Week 8
 -- -- Week 9
 -- -- Week 10
 -- -- Week 11
 -- -- Week 12 of semester 1 (Week 1 of term 2)
 -- Further background reading
 -- -- Online documentation
 -- -- Books

-- OBJECTIVES OF THIS COURSE

These objectives are not listed in chronological order.

1. To introduce general and AI-specific programming concepts and
techniques, based on Pop-11 in semester 1 (to be followed by Prolog in
semester 2). Students learn how to get knowledge about a problem into
useful form for solving the problem. This may be factual knowledge
(knowledge that something is the case) or procedural knowledge
(knowledge how to do something).

2. To support, by means of examples and/or project work the AI content
of the other courses the students are doing especially:
    Introduction to AI and Cognitive Science (SEM1A1)
    Introduction to Expert Systems (Applied AI) (SEM1A2)

3. To introduce relevant software engineering/software development
skills including:

    producing a proposal, planning a program, specifying a program,
    testing, tracing, debugging, writing reports (including user
    documentation, system documentation), reporting verbally (i.e.
    giving presentations on programs), working with other students.

4. To introduce students to a minimal useful subset of the Unix + X
window environment, and the facilities provided, including files,
directories, mail and others.

NB. The programming skills and concepts are selected because of their
potential relevance to developing explanations for complex human
capabilities and simulating those abilities in applications of AI.
This requires learning how to represent information in a useful form,
and how to specify processes that use the information.

It is not assumed that you have done any previous programming. However,
students without programming experience may have to spend extra time
practising.

NOTE: students doing the AI and Computer Science degree will
simultaneously be learning C++, using PCs instead of Unix machines.

-- TIME ALLOCATION 6.5 hours per week per module
Students will have to accept responsibility for allocating their time
between this and other courses. 6.5 hours per week is a MINIMUM
requirement for doing well on thise course.

Most of the programming work will be done in laboratories, where
research students will be availalable as ``demonstrators'' some of the
time. The language used does not run on a PC unless you run the Linux
operating system. It may be available for a low charge from the
developers.


-- SPECIFIC SYLLABUS

The course will include the following components:

-- 1. LEARNING TO USE THE POPLOG DEVELOPMENT ENVIRONMENT.

POPLOG is a sophisticated software development environment produced
originally at Sussex University and marketed internationally by Integral
Solutions Ltd. It includes several languages (POP-11, Prolog, Common
Lisp, Standard ML), an integrated general purpose text editor VED, a
graphical library, and many teaching and demonstration libraries
connected with Artificial Intelligence and Cognitive Science. The course
involves the following in the first semester.

1.a. Learning to use VED as a general purpose text editor and browser,
    and also as a mail front end and program debugging tool. This
    includes learning how to read TEACH files and HELP files.

1.b. Learning to print whole files or marked range. See HELP VED_PSPRINT

1.c. Learning about the structure of the online documentation on VED,
    Pop-11 and AI, and how to browse documentation using VED.

1.d. Function keys and ENTER commands will be used, but a "menu"
    facility will be available to reduce your learning load.

1.e. Learning to do simple text formatting using VED's built in
    commands. (See HELP RNO, HELP LATEX, HELP VED_LATEX)

1.f. Reading and sending electronic mail in the editor VED.

-- 2. INTRODUCTION TO UNIX.

Some students will also learn the following:

2.a. The concept of an operating system and its relation to a language
    system like Poplog, logging in, files, and the network.

2.b. The Unix file structure. Use of pwd, ls, mkdir, cd, chmod, and
    perhaps symbolic links so that students can easily share files.

2.c. Man(ual) files: their structure and their use. (Accessible via VED)

-- 3. PROGRAMMING AND POP-11

In the first semester the language to be used is Pop-11. This language
is very similar to the more commonly used language Lisp, in power, but
POP-11 is generally found easier to read, and to learn. In the second
semester you will learn Prolog, a logic-based language. Topics you will
learn include the following:

3.a. The idea of a programming language, with a semantics and syntax.

3.b. The main data-structures used for the course:
    Words, strings, numbers, lists, vectors, procedures,
        and possibly arrays, if there is time.
    The use of lists as patterns, and as databases.
    The implementation of list structures as chained pairs
        (See online files: TEACH LISTS, TEACH BOXES)

3.c. The main syntactic forms required for the course.
    procedure definitions
    procedure calls
    conditionals, loops, return, and abnormal loop exits.
    declarations: global and local, dynamic and lexical
    assignments to variables or contents of data structures.
    syntax for various data-types, especially numbers, words,
        strings, lists, patterns and possibly also vectors.
    See TEACH POPCORE

3.d. The use of iteration (loops) and recursion (procedures that use
    themselves as sub-procedures).

3.e. The use of the pattern matcher and the Pop-11 database
        (See TEACH MATCHES, TEACH DATABASE)

3.f. "Trace" and the Pop-11 debugger.

3.i. How to organise program files, test files, and produce
    sample output to be included in reports.

3.j. How to produce a program that solves a problem by searching for a
    solution.

3.k. A brief introduction to some concepts of grammar and parsing.

3.l. Optional extra for more advanced students:
    Object oriented programming (inheritance hierarchies and methods)
    TEACH OBJECTCLASS_EXAMPLE, TEACH ADVENT_OBJECTCLASS

-- 4. THE IDEA OF SOFTWARE ENGINEERING DISCIPLINE
This includes

4.a. Specifying what a program is going to do
    (Ontology, behavioural specification, scenario). See TEACH PROPOSALS

4.b. Writing reports on programs. See TEACH REPORTS

-- 5. AI CONCEPTS AND TECHNIQUES

5.a. The importance of representation.
    Alternative ways of representing the same information (in the
    machine or the brain). E.g. alternative ways of representing
    information about the "river" problem. Tradeoffs between
    explicitness and complexity of changes.

    General purpose vs domain-specific or problem-specific
    representations (e.g. lists as analogical representations).

    Logical representations vs "ad hoc"(?) representations

5.b. The flexibility of lists and pattern matching for representing
    and relating very varied forms of data.

5.c. Using lists to specify not only particular facts but also general
    knowledge (e.g. grammatical rules, rules to be used by a planning
    program.)

5.d. Processing strategies: the importance of search in AI.
    Different kinds of search:
    Different kinds of search:
      searching for an item that already exists, in some list or network
      searching for something that has to be created, e.g. a new plan.

5.e. Search patterns: depth first, breadth first, heuristic search

5.f. The importance of good representations in controlling search
        TEACH TOWER, TEACH SEARCHING

5.h. Parsing and grammars: TEACH  GRAMMAR

5.i. Other techniques used in AI (If there is time.)

    Use of production systems for expert systems: TEACH EXPERTS
    Use of scripts and schemata: TEACH SCHEMATA

    There are many additional teach files for ambitious students; to.

-- 6. MINI PROJECTS

The course will be assessed by means of a mini project, to be specified
later. Students will write a proposal, then do the projet and write a
report. The assessment will be based more on the written report than on
the program code. E.g. clarity of explanation, analysis of problems,
description of inadequacies in the program, etc. will all play an
important role in the assessment.

NOTE: There is a lot to learn and not everyone will cover all the
material. It will be possible for students who are new to computing to
cover less material than those with more experience. This can be
compensated for in other ways, during your three years here.

-- 7. PROVISIONAL COURSE TIMETABLE

Each week there will be set times for laboratory classes at which help
will be available.

Laboratory class times are currently arranged for Friday afternoons 2pm
onwards, in room G10, the Unix lab, on the ground floor of Aston Webb C
block (the same level as the main entrance). Lectures are Wednesdays
11am in G12.

Several times a term you will obtain report forms for specific
assignments. You will need to get a demonstrator to check your
assignment and sign the form which you will then hand in. This will
provide a check on your progress and should help students who find the
work difficult.

-- -- Week 1
Overview, followed by introduction to logging in and VED+X as an
interface

Students will start to use VED, mail, unix, playing with Eliza and
simple programming

a. Introductory VED files
    TEACH TEACH, TEACH VED, TEACH MARK, TEACH LMR, TEACH VEDPOP
    Information about keyboard maps on the NCD X terminals and Sun
    workstations can be found in
     HELP NCDXTERMKEYS, HELP SUNXTERMKEYS, HELP SUN4KEYS

b.  Play with Eliza and try to work out Eliza's rules and where they
    fail, and why.

    Students with previous programming experience try TEACH GSTART
-- -- Week 2
Lecture: on what a programming language is and how Pop-11 fits in.
    A language has syntax and semantics (virtual machine)
    Main facilities in Pop-11 and similar AI development environments:
    data-types, pattern matcher, control structures, (incremental
    compilation?), AI tools.

a. Continue developing fluency with VED, and practice using mail.

b. Use TEACH RIVER to get an introduction to problem solving and to
    notions of procedure, procedure parameters.

    TEACH DEFINE, If fluent, try TEACH GSTART

-- -- Week 3
a. Use TEACH RESPOND to develop a mini-eliza program, learning about
   conditionals, the matcher, variables.
b. Teach files on Programming. A subset of:
    TEACH DEFINE, TEACH LISTS, TEACH MATCHES, TEACH VARS, TEACH POPCORE,
    HELP IF

-- -- Week 4
a. Students who have not finished TEACH RESPOND do so.
b. Those who have, start on TEACH GRAMMAR up to the train announcement
    example. Try to produce your own grammar and lexicon for either
    railway announcements or newspaper headlines.

-- -- Week 5

    Start working on TEACH RIVER2

-- -- Week 6

    Continue TEACH RIVER2
    Advanced students can try TEACH RIVERCHAT

-- -- Week 7

    Introduction to searching
    TEACH TOWER (How to build a tower of a specified size).

-- -- Week 8

    Continue TEACH TOWER. More advanced students: TEACH SEARCHNG

-- -- Week 9

    Rule-based programming exercise. A mini expert system.

    Students start to think about mini project: TEACH PROPOSALS

-- -- Week 10

    Programming revision. Go back to earlier exercises and catch up.
    Read the Pop-11 primer.
    Work on mini-project
    Prepare a design specification (as in TEACH PROPOSALS)

-- -- Week 11

    Work on mini-project

-- -- Week 12 of semester 1 (Week 1 of term 2)

Work on mini-project (Submission date first day of next term).

-- Further background reading

-- -- Online documentation
TEACH * POPCORE
    (Printed version will be handed out)
    Very similar to appendix on Pop-11 in the Sharples et. al. book

TEACH * POPSUMMARY
    This is a more extended overview of Pop-11 (needs some updating)

TEACH * PRIMER
    This gives an extensive overview of many aspects of Pop-11. Copies
    will be made available in the laboratory and in the library, for
    browsing. A printed version can be bought for 3 pounds.

There are many online REF files, HELP files and TEACH files with
extensive documenation on Pop-11, the Editor, and the libraries.
Part of what students need to learn is how to find their way round
these files.

-- -- Books
In addition to the Pop-11 primer, there are books mentioned in the
general reading list, including:

M.Sharples et. al. (1989)
    Computers and Thought, MIT Press

The School Library has some copies of the Poplog User Guide, which can
be read in the Library, if you are feeling adventurous.

--- $poplocal/local/teach/sem1a3
--- The University of Birmingham 1995.  --------------------------------