Attitudes to APL in Higher Ed’n

Transcribed from the printed issue; not yet reviewed. Read it in the PDF of the issue, page 34

Last year the BAA commissioned a market research survey into the usage of APL in higher education. Six students at Humberside College of Further Education, in their final year of an HND in Business Studies, carried out the survey as part of course options covering Marketing Operations and Market Research.

The idea for the survey was initially discussed by David Eastwood and David Hewson who lectures at Humberside CFE and is also a BAA member.

The Terms of Reference¶

The students were given these objectives::

1. Determine the level of awareness of APL in Higher Education
2. Identify areas of usage of APL
3. Identify levels of usage for various programming languages
4. Identify hardware used
5. Gather opinions on APL

The survey was concluded towards the end of the academic year and a prsentation was made to the BAA Committee in London. The students involved were: Mark Alston, Jill Southwood, Helen Corcoran, Neil Cotton, Nadeen Vigus and Alan Clifford. Three of the Humberside CFE lecturers also attended – both to listen and to award marks to the students!

I shall outline the way in which the survey was carried out and highlight some of the results presented. The BAA holds a full copy of the report which we will make available to members on request.

The Survey¶

The survey was conducted by means of a postal questionnaire, with a local trial run, and then a mailing to 311 separate Departments in Universities, Polytechnics and Colleges of Further Education. Departments selected included Accountancy, Business Studies, Computer Science, Mathematics, Statistics and Engineering. 149 replies were received, about 60% from Universities.

Analysis of the results was, of course, carried out in APL, using a program written by David Hewson (thank you David) and the results and conclusions were presented by the students jointly. All the results were presented in graphical form on overhead foils, although I shall confine myself to reporting some of the numbers. As you will see from the various results, the questions asked were aimed at giving the BAA information about what might make more people use APL, and in this context some of the comments sent in, in reply to the final question – ‘any other comments?’, were very interesting. Some of them are quoted at the end.

Of those surveyed, 90% had heard of APL although only 35% had or did use it. This rather goes against the common feeling that more people would use APL if they knew about it. Perhaps they know the wrong things about APL? Of those who were aware of APL, most categorised themselves as ‘Novices’.

The most popular languages used were, in order: Pascal, FORTRAN, BASIC, followed by LISP, COBOL, FORTH. Others used include Prolog, C, Modula-2, Ada and Assemblers. Usage of LISP was high, with about 80 users against the 170 or so for Pascal and FORTRAN. In most cases usage was split evenly between use for teaching and non-teaching purposes. The survey enquired about why a given language was used and the five most important factors were, in order:

Relevance to an Application
Importance for teaching
Importance for research
Availability
Familiarity

A similar question was asked, to try and establish what influenced the choice of a language for teaching purposes, and here the answers were ranked:

Relevance to Course
Relevance to Industry
Ease of Use
Machine Compatibility
Versatility
Price
Availability of training packages
Statutory Course Requirement

Just to check that people actually still used programming languages, a question was put about the importance of learning a programming language. The highest response here was ‘Very Important’, with more Engineers thinking it important than other categories and more Business Studies thinking it not important to learn a computer language. Interestingly, those involved with computer studies were split evenly between the ‘Very Important’ and ‘Not Important’ ends of the spectrum.

Hardware used ranged, in order of importance:

IBM PC – BBC Micro – Vax – Prime – Apricot – Other

Conclusions¶

The students’ conclusions can be summarised as seeking to redress what is seen as an image problem for APL, and broadly they suggested:

• Simple introductory material on APL should be prepared
• Training packages relevant to specific subjects should be devised
• Publicity material should be placed in mainstream magazines and papers
• Get key decision makers on education software to attend residential courses

Other Notes and Comments¶

Given the time available for the students to find out about APL and carry out their survey, I think the results were both interesting and informative. In hindsight, the BAA could have set better questions for the survey and we are certainly planning to repeat the exercise in the near future. I should conclude by handing over to the respondents themselves (I have avoid identifying them in case they are known to readers of Vector):

“What developments would you like to see in the programming languages you use?”

– Site licences
– The ability to write programs like APL using the normal character set
– More international standards
– Few; stability is precious
– Flexibility in linking to other packages
– Cheap APL2
– Improved user interface
– Parallel implementations
– Explicit looping in APL
– More standard APL packages
– Built-in graphics

“I would like to use APL but our computing centre refused facilities”

“Major problems with APL are lack of an effective standard, unusual keyboard requirements”

(The problem of needing a special terminal is mentioned many times)

“APL should be allowed to die, we have too many computing languages”

“APL has a reputation of being a computer boffin’s language”

“APL is a dead duck because it is unreliable”

“The major problem is the strange character set and the time taken to master the language”

“How would APL benefit business computer users. Is APL available for UNIX?”

“APL encourages tricky thinking style and ignores questions of efficiency”

“APL is a horrid language. It violates most of the principles of language design. Programs are usually unmaintainable.”

“I have yet to find anyone who has learnt APL who has continued to use it once FORTRAN and BASIC were available with reasonable support.”

“Many people in APL are too ‘into’ the language to realise that the rest of the world is passing them by.”

“The operators of APL are an attractive and useful feature, but in other respects APL is very dated and lacking in features such as support for data abstraction and structured programming.”