Are 4GLs killing APL?

Transcribed from the printed issue; not yet reviewed. Read it in the PDF of the issue, page 49

Before answering this question, it is necessary to consider two related questions:

Who uses 4th generation languages?

What is a 4th generation language?

The first question is perhaps easier to answer. The concept of the Information Centre has grown over the last few years, and it is clearly in the Information Centre that 4th generation products will find their home. This gets us no closer to resolving our main topic, however, as we must still ask:

What is an Information Centre?

Why does it exist?

This discussion will first describe Information Centres, in order to understand the growth in demand for 4th generation languages. It will then use a 4th generation product, DATAPORT, to demonstrate 4th generation characteristics by example. Finally, it will show why the use of APL in 4th generation products is an advantage. It will come as no surprise, therefore, that the answer to “are 4th generation languages killing APL?” will be “NO.”

It has been suggested by sceptics (Ed: see Romilly Cocking’s article in Vol. 1. No. 1.) that Information Centres have arisen because IBM have seen the IC concept as a means of selling blue boxes. Whilst it is true that IBM have put enormous resources into promoting IC’s, it is now clear that, as usual, they have been backing a winner. During the last twenty years organisations have moved more and more of their corporate information to the data processing environment, and potential users have found it increasingly difficult to cope with the problems of accessing this information. Sometimes this has been because of the sheer volume of data available, at other times the technical difficulties of access have made the time scale of data interpretation too long. Hence the expression “application backlog” became an emotive term in many DP departments.

The Information Centre concept arose, therefore, in order to formalise methods by which non DP professionals (now often referred to as “End Users”) could access and interpret this wealth of computerised data. Professionals would be made available to advise and assist where necessary, but the users would use advanced computing tools to achieve results at terminals without the need for intervention by the DP department.

APL was introduced to the Information Centre not, as has been suggested, to fill machines, but because IBM had used two APL products — APL/DI and ADRS — internally to great effect themselves. In fact, considering IBM’s vast internal committment to APL it is surprising that APL is not pushed more into the limelight. Nevertheless, both APL/DI and ADRS found great following in emergent Information Centres, and still remain the most frequently used End User tools.

At the same time that Information Centres using mainframes as the computing vehicle started to emerge, the micro computer revolution occurred. Suddenly people found themselves confronted with computing tools which were easy to use! And productive! Visicalc changed the face of End User Computing, but this was only the beginning. Command driven systems, once considered to be at the forefront of technology, were replaced by sophisticated menu-driven systems, with on-line help facilities giving real power to the users’ fingertips, without the need for heavy committment to learning how to use the systems offered. Mainframe systems fought back, and the term “4th generation” was spawned.

“4th generation” is a term widely used, but rarely defined, since it embodies such a diversity of concepts. Rather than attempt a definition, I will use a new Information Centre product, DATAPORT, to achieve a definition by example.

DATAPORT is a 4th generation Information Centre tool providing full-screen data entry, display and analysis capabilities together with interfaces to other proprietary Information Centre products. DATAPORT uses APL as a “core technology”, draws heavily on spreadsheet experience to provide a fast, consistent environment for End User applications.

The 4th generation aspects of DATAPORT can be divided into three main areas:

user-friendliness;

functionality;

interface capabilities.

Looking first at that over-worked expression “user-friendly,” what does it mean? The experience of micro computer users is clear. It is possible to “throw” a micro and a good spreadsheet at a user and never hear from them again. Expect for a stream of converts asking for their own PC! Clearly this is a result of user-friendliness, but how is it achieved? Certainly aspects such as menu options and on-line help are part of the answer, but not by any means all. This is why older generation software does not necessarily respond to being tarted up with menus and help — a new, consistent approach is called for. This is achieved in DATAPORT by using the following:

combination of command and menu driven design;

two level help facility;

clear, consistent error messages;

full support for 3270 technology;

selection of options by cursor control;

all of which are incorporated into a truly elegant design. The result is a product which not only encourages beginners to learn without “painful” experiences, but also enables experts to work quickly and efficiently.

Friendliness is fine, but ultimately it is the results which matter. DATAPORT functionality extends the well developed concept of windowing to the multiple dimesions which are so important for modelling the real world. In combination with calculation facilities which enable formulae to be expressed in terms of rows, columns, or even pages at a time, the multi-dimensionality enables users to rapidly develop applications. It will come as no surprise to experienced APLers that the calculation facilities take full advantage of the APL product core to provide unparalleled power of expression.

It is in the interface capability, though, that DATAPORT is furthest ahead. Friendliness and functionality have developed gradually, but here is something new. The capability of “talking” to other software, whether mainframe or micro based, without the user knowing anything about the relevant data structures, is revolutionary. Current interfaces include links to database products including FOCUS and RAMIS; core Information Centre products APL/DI and ADRS; other mainframe based products including SAS and Tel-a-Graf; and simple interfacing to the popular micro spreadsheets including LOTUS and Visicalc.

Having now achieved some clarification of the implications of “4th generation” the advantages of APL in this environment are clear. Whilst the use of APL is of no direct benefit to the End User (in fact it is important that the user be unaware of APL, in terms of upper/lower case requirements, APL characters, or such “peculiarities” as right to left evaluation) the benefits for those supporting the software are tremendous. Even the simplicity of installation of APL, merely requiring the loading of a workspace, is a benefit in terms of the avoidance of the need for that all-to-scarce resource, the systems programmer. But it is in the ease of customisation that the main benefit lies. Frequently this can be achieved with minimal awareness of product internals. Even where the relevant system is not APL based, the use of APL to provide a front end, or interface to other products in the manner of DATAPOINT, is frequently because of the ease of transforming data structures in APL.

The above assumes that APL expertise is available. But the natural environment for 4th generation products is in the Information Centre, with its historical use of APL. It may well be that APL is the only language supported by an IC; in any case IC’s will be well equipped to work with APL products.

It is seen then, that rather than killing APL, the emergence of 4th generation languages, whether APL based or not, will stimulate the continuing growth of APL as the vehicle for the best of End User software.