courseware - software interfaces: some designs and some

advertisement
COURSEWARE - SOFTWARE INTERFACES: SOME DESIGNS AND
SOME PROBLEMS
Ms. D. M. Paramskas
Abstract
When designing a free-form remedial/tutorial program for second language learning, several principles are essential to insure
pedagogically sound courseware. These include adaptability in terms of instructor requirements and user population needs; user
friendliness to reduce machine-induced phobias and compensate for mechanical ignorance; sophistication in error analysis through
extensive branching in order to produce comments which will help the user to learn through his/her errors; and flexibility, which
gives the user control over both the content and the pace of learning. There are some disadvantages to such a program, but newer
projects aim to overcome these disadvantages by allowing the individual instructor more freedom in creating materials for specific
needs. Finally, evaluation through statistical records is discussed as a point of contention.
To describe CAI courseware, as opposed to demonstrating it, can be done, but like describing music rather than playing it, is very
difficult to do well. The purpose of this paper is to set out the general plan of one system, to raise some questions, in the hope that
inevitable ambiguities will be resolved during discussion.
CLEF1 is a courseware system designed for remedial/tutorial practice at the beginning level of second language learning (in this case,
French). It is composed of 62 lessons covering basic French grammar. Each lesson is made up of 1) a brief grammar explanation of
the grammar point to be practiced; 2) vocabulary drill for words used in the exercises which might not be transparent to a typical
Canadian anglophone; and 3) a set of 3-5 exercises.
The designers of CLEF were particularly interested in creating a system which would be adaptable to a broad range of texts and
methods, user-friendly, sophisticated in terms of error correction and computer response, and flexible enough so that students with a
broad range of skills could equally profit from using it.
Adaptability
Each lesson is contained on one diskette; each lesson drills one grammar point only. Instructors are therefore able to shuffle the order
of the lessons to suit whatever text or method they are using. The vocabulary used relies on cognates and sets of words typically
found in beginning-level texts. When other kinds of words occur, they are first presented in an introductory vocabulary drill. As a
fail-safe, and to deal with a user population other than the target one (as, for example, students whose first language is not English and
for whom transparency of vocabulary is not evident) there is a dictionary function within each exercise. Users may at any point type
"?" plus the word they do not understand. The translation appears at the bottom of the screen.
User-friendly
The program is activated by inserting the diskette and pushing one button. Simple instructions then appear on the screen, including a
request for the user's name, which then occurs in the responses as in "Sorry, NAME, try again" or "Congratulations, NAME."
No knowledge of computing is required, and two-finger typing skill is adequate for use. If lesson material is limited to one side of the
diskette, users may insert either side as the same lesson is duplicated on both. If lesson material is longer, users may insert either side,
and when requesting an exercise which occurs on the other side, they will see the flag: "For exercise X, flip disk and press RETURN."
Names of users are stored in memory and they may proceed without any difficulty or need to re-insert identification.
Sophistication
Computer responses to user error are a key factor in any courseware. They must be able to anticipate common errors and give clues
specifically designed to help users overcome that particular error. CLEF avoids as much as possible catchall responses such as "That's
not right, try again."
CALICO Journal, Volume 1 Number 3
4
Errors have been classified into three categories: flubs, non-significant errors, and significant errors. Flubs are the result of spastic
fingers - extra spaces, trailing spaces, lack of spaces, and odd capitals. These are disregarded if the answer is correct, and are
automatically rewritten. Non-significant errors include single missing letters, double letters, and single extra letters. The response to
these is "there is(are) extra letter(s)," "There is (are) letter(s) missing" or, simply, "A typing error?"2 Users are allowed two tries for
each question, but non-significant errors are not counted as tries.
Significant errors are those which touch upon the grammar point being drilled (i.e., verb endings, adjective agreements, etc.). In this
case, anticipated common errors each have an appropriate message i.e., "Attention, NAME, the subject is plural, not singular," "The
ending is right but the stem is wrong," etc.). Zero or garbage input produces a catchall "Sorry, not correct, try again." If the correct
answer is not given after the second try, the answer is printed on the screen, and the next question presented automatically. The nonanswered question reoccurs in a "revision" at the end of the exercise.
Messages for correct answers are varied, random and two-tiered, ranging from Gallic excess for a correct answer at first try
("Fantastique!") to a subdued "OK" on the second try.
In some cases, as in exercises matching articles and nouns, users may vary the article displayed before choosing a noun. Whenever
possible, alternate answers are accepted as correct. For example, in a lesson on time, 6:15 may be acceptable given as six heures et
quart, six heures quinze, dix-huit heures et quart, dis-huit heures quinze. Exercises may be chosen from the menu listing in any order,
may be repeated or skipped, and the user may exit exercise at any point and be returned to the menu.
Flexibility
Learner motivation is an extremely important factor in the success of language acquisition. Motivation generally decreases where the
learner perceives himself as caught up in an inflexible system whose pace and feedback are determined by a "machine". The pattern
drills which use the global-structuralist approaches are good examples of demotivating, mechanical formulas. Motivation generally
increases where the student sees himself in command of his learning, and has choices in pacing and access. CLEF attempts to favor
high motivation both by giving control of pace to the learner, and through a series of options.
Early CAI materials often relied on an internal time to guide learners: so many seconds for viewing this screen, so many seconds to
arrive at the right answers, etc. It was felt that speed of acquisition was a factor in learning language, and deviation from a standard
speed, regardless of right answers, indicated some sort of flaw. The CLEF system rarely uses a timer, but rather depends on a footnote
("Press RETURN to continue") for pacing. In effect, this means that the fast learner and the slow learner may each go at his own
speed, without either feeling penalized, rushed or bored.
In addition, a series of options were developed to reinforce user ability to organize pacing and access to materials to suit his own best
learning situation. Options include the translation and "quit" options mentioned above. Within each exercise, users may recall a
summary of the grammar in point form - in more complex exercises, students have been known to forget just what was important at
the moment. In fact, it is not completely unknown for students to come straight out of class with only a hazy notion of what the
grammar point explained was all about. CLEF includes the extended grammar review at the beginning (it is optional) for the latter,
and the point form review, also optional, for the former.
Vocabulary drills are included at the beginning of each exercise but, again, instructors are familiar with student talent for suddenly
forgetting not only new words, but ones which it is blithely assumed they must surely know by now. The translation option previously
described is the remedy. All these options reinforce a basic tenet of language instruction: go from the known to the unknown, and
learn one point at a time. Students having difficulty both with meaning and grammar are not likely to profit much from an exercise. If
meaning is the important point, then grammar help should be readily available; if grammar is the central point, then easy access to
meaning will create a more relaxed learning situation.
Future Developments
The designers feel that CLEF has met all of their proposed objectives. However, there is one drawback: the system is rigid, as any
printed textbook can also be called rigid. Unless instructors are programmers or have access to experts, they cannot change any item
or insert materials they would like to drill; and even if a programmer were available, the amount of time needed to program and debug
would be prohibitive. Plus, all instructors know that no system ever completely meets their needs unless they themselves have
designed it.
To meet these objections and create even more flexible materials, several projects are now under way. One of these is a programmer's
authoring system for drill-type exercises,3 in which correction routines from CLEF are written so that instructors may prepare content
and specify the type of correction desired (branching, hierarchies of response, spelling checks, etc.). Any programmer can then
implement the instructions in a minimal amount of time.
A second project dispenses with the programmer completely, and aims to create a template4 (or shell, or framework program) for
specific types of exercises (multiple choice for comprehension purposes, fill-in blanks for either comprehension or grammar,
dictation). Instructors are prompted to insert their materials, which are then automatically formatted. Once the input is complete and
CALICO Journal, Volume 1 Number 3
5
edited, the template automatically churns out the student version.
Conclusion
There seems to be developing a general consensus that CAI in second language learning is best used in the tutorial-remedial mode.5
But evaluation of the effectiveness of CAI in general is an ongoing problem.
It will be noted that the whole question of evaluation (through statistics stored for retrieval by instructors, as well as pre- and posttests) has been thoroughly avoided up to now in this paper.6 This type of evaluation has become widespread in CAI materials through
what seems to be, in my opinion, an unfortunate misunderstanding. Early CAI materials were developed for science, and were
generally designed as modules for independent study. Measurement of learning in the sciences, math and computing, deals with the
quantity of information acquired, and this information often comes in terms of black and white, right or wrong. Language learning,
indeed learning in all of the humanities, cannot or rather should not be measured in similar terms. The fact that so many students have
completed so many exercises with so many right/wrong answers does not guarantee that those students have learned how to use the
language, which, after all, is the primary objective. At best, it tells instructors that students have been exposed to a certain amount of
practice, and testing that exposure should be a function of using the language in realistic situations, not in isolated phrases with highly
structured sequences. Drill is needed, but it cannot be quantitatively evaluated. Computerized practices just one of the colors in a
rainbow of learning opportunities. For reinforcement and remediation in grammar, and practice in written comprehension, it is an
excellent tool, but only one of many. And, of course, the tool is only as effective as its content. No matter how sophisticated and
spectacular the software, courseware remains the key factor.
NOTES
1
CELF(Computer Assisted Learning Exercises for French) is a joint project of the Universities of Western Ontario, Guelph and Calgary.
Additional documentation can be obtained from the Computer Program Coordinator, Faculty of Arts, Language Laboratories, University of Western
Ontario, London, Ontario N6A 3K7.
2
This kind of spelling check is a good example of the advantages of cooperation between instructors and programmers. For instance, a
programmer was told to write a spelling check routine for an exercise on time in which up to four answers could be acceptable and all were made up
of three words, as in the previously cited 6:15 - six heures quinze, six heures et quart, dix-huit heures quinez, dix-huit heures et quart. The
programmer with no knowledge of language teaching or learning, assumed that mistakes in any part of the answer were significant. After many long
weary hours, indeed weeks, she devised a brilliant routine. Its only drawback was an excruciatingly slow response time, up to 10 seconds in one
case. Clearly, this was not acceptable. When the lesson arrived for debugging, it was immediately obvious to me, as a teacher, that only the first and
last parts of the response were significant. Any errors occurring in the central heures part could only be flubs. Therefore, any errors in the central
part were branched to the response "A typing error?" and checks for the other two parts could be made within a reasonable response time.
3
A project of the University of Western Ontario.
4
A joint project of the University of Guelph and University of Calgary.
5
Solveig Olsen, "The Prospects of CAI in Foreign Language Studies: in MICRO 1,3 (September, 1980) p. 6.
6
While still a minority view, a sort of anti-statistics movement seems to be building in CAI. See S. K. Lower, G. Gerhold and L. Braun's
comments in "Evaluation 1" in K. Lewis and E. D. Tagg, eds., Computer-Assisted Learning: Scope, Progress, Limits (Amsterdam: North Holland
Publishing Co.) 1980, pp. 201-101. Aiken and Braun sum up the position as follows: "Various statistical techniques have been used to measure the
effectiveness of CAI materials. Pretest and posttest scores for example, have been used along with other traditional methods. However, since so
many assumptions are made and so many variables are uncontrolled (teacher differences among the most important of the latter), such statistical
results have been meaningful only as measures of performance; other methods will have to be considered if we are to have meaningful measures of
learning." R. M. Aiken and Ludwig Braun, "Into the 80's with Microcomputer-Based Learning" in Computer 13, 7 (July 1980) p. 13.
CLEF Bibliography
Harley, R. E., H. Heller and G. Holmes “Input Analysis and Feedback in a CAI SYSTEM for Languages,” in 1981 Conference Proceedings of the
Association for the Development of Computer-Based Instructional Systems (ADCIS): Computer-Based Instruction: Frontiers of Thought.
Computer Centre, Western Washington University: Bellingham, WA, 98225 U.S.A.
Kidd, M.E. and G. Holmes, “Realizing the Potential of the Computer in the Learning of Modern Languages,” in Proceedings of the Third Canadian
Symposium on International Technology, National Research Council Canada. Ottawa. 1980.
Holmes, G. and M. Kidd, “Servicing Learner Needs: From Teletype to Micro,” in Proceedings of the 11th Ontario Universities Computing
Conference: Interface, the Challenge of the Eighties. Queen’s University: Kingston, Ont., 1980. Also to appear in System. 1981.
Kidd, M. E. and G. Holmes, “CAI for Language Remediation at the University Level,” in Proceedings of the Western Educational Computing
Conference 1980. California Educational Computing Consortium: 1982.
Holmes, Glyn and Marilyn Kidd, “The Evolving Case for Computers in the Study of Modern Languages” in ALLC (Association for Literary and
Linguistic Computing) Journal I, 1, pp. 7-11.
Paramskas, D. M. and Nicole Rolland-Vassiliadis, “An Evaluation of the Effectiveness of Micro-computer based CAI for Second Language
Teaching,” paper read at the Sixth International Congress of Applied Linguistics (AILA) Lund, Sweden, 1981.
Holmes, Glyn and Marilyn Kidd, “Computer-assisted Learning: Design and Implementation,” in SPEAQ Journal IV, 3-4, 1980. pp. 83-96.
Mydlarski, D. “So you Want to Do Your Own CAI Program; the Experience at the University of Calgary” in ibid, pp. 119-128.
Paramskas, D. M. “CLEF: A CAI Project for Elementary French”: Paper read at the 1982 ECOO Conference, Toronto, Ontario. May 1982.
Holmes, Glyn and Marilyn Kidd, “The CLEF Project: Learning French on Color Micros” in Conference Proceedings, ADCIS. Bellingham: Western
Washington University, 1982. pp. 245-251.
Holmes, Glyn, Donna Mydlarski and D. M. Paramskas, “Co-operative Courseware Production: A Solution to the Problems of CAI?” in ibid, pp. 252258.
CALICO Journal, Volume 1 Number 3
6
Download