Skip to content

Latest commit

 

History

History
170 lines (81 loc) · 8.09 KB

meet3_notes.markdown

File metadata and controls

170 lines (81 loc) · 8.09 KB

APL ∊ BCN - Meet #3

Historical APL paper reading group

Contents

  1. Structure & goal of the historical APL paper reading meetings
  2. The paper
    1. Summary
    2. Introduction
    3. The Epidimiology of APL
    4. An Annotated Bibliography
    5. APL Implementations
    6. APL Compatible Terminals
    7. APL File Handling Capabilities
    8. APL Handbooks
    9. Conclusions
    10. Bibliography
  3. Discussion & questions

Structure & goal of the historical APL paper reading meetings

Structure

  • Beforehand you'll be given the paper of the forthcoming meeting so you can have a fast look.
  • Many papers will be taken from the JSoftware website where they have a nice collection of important and historical APL/J papers.

Goal

  • Get a deeper understanding of the founding languages of the array programming paradigm (APL, J).
  • As a consequence of the previous goal, become better APLjk programmers.

The paper

IS APL EPIDEMIC ? OR A STUDY OF ITS GROWTH THROUGH AN EXTENDED BIBLIOGRAPHY

J. C. Rault and G. Demars, 1972-01-01

Excerpts taken from the paper.

a. Summary

An attempt is made to demostrate that the use of APL is growing in an epidemic fashion. The theory of epidemic processes is applied in an approximate manner by means of data provided by a nearly-exhaustive bibliography given as an appendix. APL is proved to be undoubtedly epidemic.

b. Introduction

The few conferences held these past years on APL have demostrated a fast growth of its use. At this time APL, as we hope this conference will show, is pervading every area of activity.

To the dismay of the vilifiers of APL, this pattern of development resembles the spread of an infectious disease.

The purpose of this paper is twofold: first establishing that APL is an epidemic, a commonplace assertion among APL supporters; second providing the APL community and the APL addits to-come with an extended bibliography given in appendix. To our knowledge, such a bibliography has not been made available so far.

c. The Epidimiology of APL

The spread of scientific ideas has already been studied in terms of an epidemic process [1, 2, 3]. A thorough theory has been laid down and applied to the study of an entire discipline such as symbolic logic [3] for a one-hundred year time interval.

Along these lines we attempt here to apply the same theory to the growth of APL using as a database a bibliography recently compiled.

APL_contributors_py_vs_epidemic_cure

Distribution_of_pubs_vs_ratio_of_pubs

It would not be contended that this literature search is exhaustive and we are fully aware that improvements can be made.

This bibliography amounts to about 330 entries and 220 authors. Thus the APL epidemic process, investigated here, considers a population of 220 individuals, or infections, over a ten-year span. Taking into account that an entry may include several authors the total number of publications is 422.

Figure 1 shows the number of new contributions each year; figure 2, showing the change in the number of active contributors each year, represents the epidemic curve for APL since its inception. This curve reveals clearly that since 1968 APL is really an epidemic. This corresponds to the release of an "infectious material" by IBM, APL/360 as a class III product.

The shape of this epidemic curve does not allow one to foresee that the epidemic will stabilize in the near future, which, according to theory, should occur when the curve ceases to increase. (In fact this curve tends to grow exponentially).

Figure 3 gives the yearly number of publications; if the present rate of growth is maintained this year we may expect 250 papers in 1972 with 40 new authors.

Figure 4 indicates the yearly average number of publications per author. This ratio has increased continuously since 1969.

The above figures should not be taken as accurate ones but just as mere benchmarks manifesting that APL, in spite of its infectious character to certain people, is hale and hearty and thriving at a pace which may endanger soon the bailiwicks of those die-hard fossils that FORTRAN, BASIC and other patters are.

d. An Annotated Bibliography

Perusing the appended bibliography is sufficient to be convinced that APL is present in many fields. We intend here to make general comments for facilitating the use of this bibliography.

e. APL Implementations

One may note the fact, which is not always well known except to specialists that APL may now be found on major computers outside of IBM:

  • BURROUGHS: 23, 98, 144-46, 247, 298
  • CDC: 58
  • CII: 157-8, 181-2, 195-6, 198
  • DEC: 216
  • HONEYWELL/GE: 96
  • IBM: 22, 59, 77-80, 82, 312, 331
  • UNIVAC: 339
  • XDS: 21, 242-43, 285, 340
  • Microprogrammed APL machines: 16, 45, 102-106, 203, 207, 295-297

f. APL Compatible Terminals

A wide range of APL compatible peripherals are available. Reference 177 gives a nearly complete list of them (48, 62, 86, 101, 188, 250, 307).

g. APL Handling Capabilities

Users of APL quite early have demanded facilities to work with large collections of data under program control. A number of file systems have been experimented or are presently available (37, 63, 84, 179, 189, 202, 262, 334, 338).

h. APL Handbooks

Many books, handbooks, user's manuals, videotapes and other materials are available: 6, 10, 23-24, 42, 78-80, 95, 112, 140-2, 205, 219, 232, 235-38, 320-323, 329.

APL_materials

i. Conclusions

Aside from the polemical aspect of this paper aimed to pique APL detractors, hope that it will be a contribution to the spreading of APL. We propose that this bibliography to be augmented, improved and refined, possibly with the help of a KWIC index.

In the following we give for the most significative fields of application, the appropriate references where the reader may find more details:

APL_applications

Bibliography

  1. W. Goffman. "Mathematical approach to the spread of scientific ideas", Nature, 212, pp. 449-452, October 1966, DOI.

  2. W. Goffman and V. A. Newill. "Communications and epidemic processes", Proc. Royal Soc., A298, pp. 316-334, May 1967, DOI, JSTOR.

  3. W. Goffman. "A mathematical method for analyzing the growth of a scientific discipline", Journal of the ACM, Vol. 18, No. 2, pp. 173-185, April 1971, DOI.

Discussion & questions

Questions

  • Why do you think that APL has stopped being an epidemic as the article claims that it used to be from 1962 to 1972? If APL was an epidemic in the 70s what do you think was the "cure"?
  • Which programming languages today you think could be thought to be spreading as fast as an epidemic?

Future work

  • Do a bibliometric analysis of 10 most popular programming languages of today using an array programming language and compare the results of that analysis with an epidemic model over a population of equal size. Try to see if it's thus possible to reach similar conclusions for any of the 10 most popular languages of today.
  • As a continuation of this study, perhaps do the same for the 50 most popular languages of today; As a next step, see how epidemic-like are any of the functional programming languages of today (Haskell, Erlang, Elixir, Clojure, Scala, Kotlin, etc).
  • Reproduce the results of this paper by digitizing the APPENDIX of the paper and plotting the bibliometric results. Also compare bibliometric results with various infectious disease spreading models.