From Wikipedia, the free encyclopedia
Jump to navigation Jump to search

Current total for lines of COBOL programs[edit]

The article (based on 1981 data?) claims that little new code is being written in Cobol. A more current estimate is at 5 billion codelines a year, so perhaps it depends on the definition of "little"... (See for instance, citing Gartner Group as a source.) — Preceding unsigned comment added by (talk) 20:05, 9 February 2004‎

Split History section into separate article[edit]

The article is currently 109 kB long. The size guildline WP:SIZERULE says an article with prose over 100 kB "almost certainly should be divided". The essential part of this article is the exposition of COBOL's features and I think the "Criticism and Defense" section gives a useful (but slightly less essential) overview of the language's flaws and reaction to it. So, that leaves the History section looking disproportionately large.

The History section has enough sources and is long enough to justify being its own article. Viz., it relies on the HOPL I conference proceedings (which had an entire session dedicated to COBOL), Kurt Beyer's biography of Grace Hopper (which focusses on her role in the DoD design phase), Jean Sammet's history of COBOL and a miscellany of articles from Computerworld and other figures involved in the design of COBOL. The History section takes around 30 kB, so removing it and replacing it with a 5 kB summary should get us acceptably close to WP:SIZEGUIDE's "Probably should be divided" range.

I am open to a more aggressive removal of content, though I can only see the History section as a good place to do so. In particular, the "Background" and "COBOL 60" sections could be shortened and the sections from "COBOL-74" onwards could have the standards' lists of features better summarised.

EdwardH (talk) 20:24, 17 July 2018 (UTC)

I think the "Features"/syntax section is also excessive, and might be a good section to split into its own article COBOL syntax. power~enwiki (π, ν) 18:13, 19 August 2018 (UTC)
@Power~enwiki: I think a COBOL syntax would come a bit too close to infringing WP:NOTMANUAL and would struggle to have enough references for notability. We could reduce the Features section anyway. I feel the historical "Hello, World!" example is unnecessary - it's more about MVS and JCL than COBOL. Uninteresting obscure features (e.g. 66 and 77 level-numbers) could be removed and the overview of statements could be compressed. EdwardH (talk) 09:33, 25 August 2018 (UTC)
The COBOL language is verbose, its programs are correspondingly long. Why should an article about COBOL be short?
That's a joke, but seriously speaking, COBOL is important for its history, and because there are still alive many inherited systems written in COBOL. For that reason, it is not a bad idea to keep all together.
It would be better to rewrite it more concisely, but not removing "uninteresting obscure features", contrary to that, it would be better to compare those features with contemporary languages or an abstract language.
For example: How does a variant or union type in C or Pascal, written in COBOL with the 88 level?
That is very important, because so, this article could cover, both, the features of the language and the anecdotes around this language, both are part of its history.
The belief that the 66 or 77 are "obscure and uninteresting" is shared by many programmers, but it is important to keep them, because they are related with low-level features, COBOL, in some way, seems a verbose assembly language. — Preceding unsigned comment added by 2806:107E:C:F09:218:DEFF:FE2B:1215 (talk) 11:33, 15 September 2018 (UTC)

Section "Influences on other languages"[edit]

Is there information about COBOL's influence on Perl? There are no details at [1] (which mentions COBOL, without any detail, among the languages from which Perl features were taken) and [2] (Tom Christiansen?) only says that Perl wysiwiggery comes from COBOL pictures. Apokrif (talk) 23:21, 4 September 2018 (UTC)

COBOL pictures [3] appeared in Perl, but they're pretty trivial. Andy Dingley (talk) 00:20, 5 September 2018 (UTC)
Did really COBOL influenced Pascal and other languages with its record and variant types? NO!
It is often believed that Pascal records (the hierarchical nested levels) and the 88 level and redefines for variant records, influenced Pascal, (I don't remember the details of how, that is why I came to this article) that is not true, those structures are very old in mathematics, they are called: product and disjoint union. Moreover, there is a talk about women and computing or something like that in Youtube, given by a woman who was a collaborator of Grace Hooper. There she told that one collaborator who worked for Rand, had a language with the records and variant types, which was the only thing that Hooper didn't invent, "all the rest was created by Hooper".
That reflected that the main concern in the design of COBOL, was all the bureaucratic aspects, like who wrote what and who ordered it, and so on. They didn't understood that a program is a mathematical proof, and as such it can be systematically developed with the problem solving methods and heuristics. They naively thought that programs were cooking recipes, a very widespread metaphor that encouraged many people to became programmers. Programming is a more serious activity which requires a good background in computing science. Sounds like a pleonasm, but many programmers just learn the syntax of a language and start to write code. Was that inherited from COBOL? I think that it is true in the extent that COBOL promoted programming in English, not in "weird" mathematical notation.
This article could compare critically, what good and bad features does COBOL have, on the light of an abstract well founded programming language or comparing with well designed languages. — Preceding unsigned comment added by 2806:107E:C:F09:218:DEFF:FE2B:1215 (talk) 12:52, 15 September 2018 (UTC)

Backus–Naur Clarification Requested[edit]

The text makes two differing statements about Backus-Naur:

Although Backus–Naur form did exist at the time, the committee had not heard of it

rather than the new Backus–Naur form because few committee members had heard of it.

And it is possible that neither is accurate - the wikipedia article on Backus-Naur says it was first published in the ALGOL-60 report. COBOL's first version was 1959. So it may not have been available to be heard of, like Unix at the time. — Preceding unsigned comment added by (talk) 08:31, 9 September 2018 (UTC)

Thanks for pointing this out. Looking at Sammet 1978b, the correct sentence is "the committee had not heard of it":

Certainly, BNF as a concept was not known to us. Again, remember, John's paper was only being presented in June; we were meeting in June, and I don't think any of us even saw the paper thereafter. [This refers to Backus (1959). See the references in the formal paper--ed.] When COBOL was defined it was severely criticized because the definition was not BNF.

I can't tell whether it was possible for anyone on the committee to have heard of it. Backus presented his work in a 1959 conference in Zurich and I don't when the proceedings were published or if pre-prints were shared. Backus worked at IBM as did a number of committee members. Criticism for not using BNF definitely came after the report was published in 1960. EdwardH (talk) 10:35, 9 September 2018 (UTC)