Talk:Timeline of programming languages/Archive 1

Analytical engine
The analytical engine shouldn't be on this list either, I don't think. It was a computational machine, but it did not run on a pgoramming language, rather on punchcards. Also the machine wasn't never tested until 1991. Subversive 07:40, 28 October 2005 (UTC)

While it's true that the Analytical Engine was a computer rather than a programming language, it did have a machine language which is what Ada Lovelace used to write her program. As for the punchcards, I fail to see your point. All computer programs had to be punched onto paper tape or punch cards until the mid-60s. In fact when I first learned to program in the mid-70s I was still using punch cards for my FORTRAN code. Each punch card had one language statement punched into it. This was just as true for the Analytical Engine as it was for the CDC 6600 which I last fed punched cards to. -- Derek Ross | Talk 04:58, 29 October 2005 (UTC)

Am making a second section for items where the date is unknown or questionable. -- Buz Cory

I see you moved the legend back to the bottom of the page, and that's fine, but I think the link back to the programming language page should have been kept. --loh
 * As a matter of fact, I was forced to reformat some (select/paste from NS6 to my editor lost the leading spaces) and so put a new legend at the top. You're right about the programming language link, it is now at the top. I only left the old legend at the bottom because I thought someone might want to make something of it. --Buz Cory
 * As a matter of fact, I was forced to reformat some (select/paste from NS6 to my editor lost the leading spaces) and so put a new legend at the top. You're right about the programming language link, it is now at the top. I only left the old legend at the bottom because I thought someone might want to make something of it. --Buz Cory

I agree that there is far too many versions of the same things. If we are going to include all languages on this time line it will be unusable. Really it should just show the major movements in the programming world.
 * Actually, if only languages are mentioned, there are several hundred, probably (including all variations). Not an unworkable number. There is a separate hardware timeline page, and separate timelines for OSen, and another for Commercial End User Apps might be a good idea.
 * Actually, if only languages are mentioned, there are several hundred, probably (including all variations). Not an unworkable number. There is a separate hardware timeline page, and separate timelines for OSen, and another for Commercial End User Apps might be a good idea.


 * Just don't mix them all on one page! --Buz Cory

I think you are out by an order of magnitude for the number of languages. Now I don't dispute each of these languages deserve a page on the wikipedia, but by having a time line consisting of x000 languages is not really going to convery much useful info.

What I may get around to is for the Lisp family, just show LISP, Common-Lisp, and Scheme in the main time-line but ALSO have a LISP-time-line. Other candidates would be FORTRAN-like, C-like, dBase-like, Pascal-like languages etc.

Any views on this approach?
 * If you are right on numbers, your approach is probably better. For the moment, I think this page, as it is, is probably best. If/when it gets too big, we can always split it. --Buz Cory
 * If you are right on numbers, your approach is probably better. For the moment, I think this page, as it is, is probably best. If/when it gets too big, we can always split it. --Buz Cory

I noticed that some entries here are at variance with the corresponding ones at History of computing. I have no way of deciding these, but I think they should be mentioned and eventually, we should try to resolve these. --AxelBoldt

Language      Programming Timeline         Computing History ============================================================= FORTRAN          59                           54-57 LISP            59                           58-60 Algol           58                           60 APL             62                           61 COBOL           60                           59-61 Turbo Pascal    83                           84 Ada             83                           79

Lisp: as explained in LISP_programming_language, McCarthy claims to have invented it in 1958, he makes this claim in his 1960 paper. It is doubtful that any implementation actually existed at the time (but that is true of many languages). --drj What date do we use in the Programming language timeline: the date of first implementation or of first description? --AxelBoldt

Well, I would say first implementation or first description. There are plenty of languages in which the implementation came first (perl, python, C) and plenty of languages in which the description came first (Lisp, Algol?, CPL, intercal). Reading the papers that the creators of these languages wrote one gets the impression that if the implementation came first then that was when the language was created and if the description came first then that was when the language was created. Which all seems pretty sensible to me.

Languages that evolve from others, for example B to NB to C, are more problematic because there was probably a continuous series of compilers that grew away from language and towards another. Even once the bootstrapping stage is reached.

Intercal is an interesting case as well because the language existed for 8 years before anyone wrote an implementation of it. Similarly it is not clear to me whether CPL ever had an implementation. In what sense are these computer programming languages? The facts seem to indicate that computer programming languages are more uses than the mere programming of computers, they can be used to express ideas between programmers or mathematicians for example (indeed, programs are speech!), so it isn't even necessary to have an implementation to be called a computer programming language. That was a bit more than I intended to say really. --drj
 * Even when a language hasn't been implemented it may still be very influential. CPL led to BCPL which led to B which led to C which led to C++, therefore CPL is well worth recording.  CPL may well have been implemented in the early 1970s as a student project but it doesn't really matter.  The ISWIM language has never been implemented as far as I know, yet the modern functional languages such as ML, Haskell, Miranda, etc. owe a lot to it, so again it is well worth describing. -- Derek Ross

RPG is missing in the timeline! PHP ?

C++ and C with Classes were in "unknown or questionable dates" but need not to since Stroustrup describes the years pretty accurately on his home pages. So I moved them to the normal timeline.

Hard to believe that 1975 Altair has had Steve Allen as a co-author since the first iteration of this article. (BASIC would make a boring talk-show) Dmsar 21:31 23 Jul 2003 (UTC)

Java definitely has predecessors. Its line of descent is Algol > BCPL > C > C++ You only need to look at its syntax to see that. Why remove C++ ? -- Derek Ross 17:33, 20 Feb 2004 (UTC)

The year headings ("pre 1950s", "1950s", "1960s",...) shouldn't have colspan="3" when there are now 4 columns. For forward compatibiliy, make it colspan="0", ie span all columns. (It's in the W3C standard: http://www.w3.org/TR/REC-html40/struct/tables.html#adef-colspan )

FORTRAN IV is listed as originating both in 1962 and 1966. I wasn't around back then so can not claim to be an expert, but all my information suggests 1962. Perhaps the entry in 1966 was meant to be FORTRAN 66? Am new at wikipedia and will attempt to update it if I don't hear back.

Here's a reference which supports my claim re changing it to say FORTRAN 66 instead of FORTRAN IV in 1966: http://www.ibiblio.org/pub/languages/fortran/ch1-1.html JMCorey 22:59, 4 Mar 2005 (UTC)

Lambda calculus
I have removed the lambda calculus from the list.

The lambda calculus as originally defined and used by Church was not a programming language. It was a notation for clarifying the semantics of free and bound variables. Church's interesting (and surprising) result was that this very simple system can support full arithmetic. It was John McCarthy who took the lambda notation and used it for anonymous functions in Lisp, though he did not actually exploit (or understand!) its full power, which was only done later by Sussman and Steele. If the lambda calculus is going to be in here, then all sorts of other mathematical systems should be, too, including: combinatory logic, finite state automata, regular expressions, and even first-order logic and&mdash;why not?&mdash;the conventional mathematical notation of functions and operators! All of these have been used as bases for programming languages. --Macrakis 7 July 2005 13:58 (UTC)

PHP
Our PHP page includes a start date of 1994, and an author of Rasmus Lerdorf, rather than PHP3/1997/unknown as listed in this timeline. Any idea why? Ojw 13:23, 17 July 2005 (UTC)
 * I believe the PHP page is right. But PHP 3 is a rewrite, in 1997, I have added author. Splang 07:06, 13 May 2006 (UTC)

Inclusion of languages
I think Fortran 2003 should be included, because it shows how one of the oldest languages is still being developed and in active use. --R.Koot 13:40, 14 August 2005 (UTC)

That's true. This timeline does have a place for minor languages which were very influential or which formed part of the ancestry of major ones: hence Cowsel, ISWIM, and CPL. It also has a place for current development of major languages. But that doesn't mean that it should include all the minor languages ever invented. To do so would lead to a forest/trees problem. -- Derek Ross | Talk 05:30, 15 August 2005 (UTC)
 * Could you add Fortran back to the list, please. I don't want to get banned for a 3RR. --R.Koot 13:59, 15 August 2005 (UTC)

About Chrome
Apparently, this is more a compiler than a programming language. I request advices about that. Splang

Object-C in 1983 not 1982
It seems that Object-C was released in 1983: http://www.oreilly.com/pub/a/oreilly/news/languageposter_0504.html

Fortress
Other quite important langauges: Fortress (new Fortran + Java for scientific calculations), Maude (logical and functional programing), Sather and pSather (pure object oriented language and big iterators). —Preceding unsigned comment added by 149.156.67.102 (talk) 14:20, 9 September 2007 (UTC)

Odd ideas of "historically important programming languages"
I am highly skeptical of a number of the languages listed from 1995 up to the present, particularly the really recent ones. Determining the historical importance of a language is difficult when it's only existed for a few years. Furthermore, there are a number of unusual entries which don't seem to fit - PIKT (1998) is an IDS, not a programming language; Cω is an extension to C#, not a language of its own; and there are separate entries for each revision of various languages (C, Fortran, and ADA are particular offenders).

What's needed urgently here is a definition of what we mean by "historically important". Right now, there appears to be a problem with new languages getting added indiscriminately, which ruins the utility of the timeline. Zetawoof(&zeta;) 22:06, 10 March 2008 (UTC)

Well there's a few ways that significance could be defined. One reasonable definition of historically significant could be gained by counting how many other languages it has influenced. That would rule out most of the "I just invented it" entries but still allow a place for highly influential but unimplemented languages like ISWIM or CPL. We don't need to set the threshold very high. If it influenced one or two later languages that would be enough. -- Derek Ross | Talk 18:46, 28 June 2008 (UTC)
 * If it doesn't have its own Wikipedia article, it's probably not notable enough, so you could set the threshold there.--Fiskjuice (talk) 20:34, 17 October 2009 (UTC)


 * True -- Derek Ross | Talk'' 01:51, 18 October 2009 (UTC)

Seed7
The article about the Seed7 programming language has been deleted. I think we can remove it from this timeline as well. Ghettoblaster (talk) 12:39, 28 June 2008 (UTC)
 * Done. List is still a mess (I need to get to cleaning it up someday), but marginally less so. Zetawoof(&zeta;) 07:13, 29 June 2008 (UTC)
 * Having an article and being on this list are two different things. There are languages which do not get an article, but still deserve to be mentioned at some places. Hans Bauer (talk) 09:09, 29 June 2008 (UTC)
 * For languages of historical value which we just haven't written about yet, I can see that. However, this is a language which we specifically opted to not have an article on a few years ago, because it had very few users and no influence on other programming languages. For these exact reasons, it's also inappropriate in a timeline of programming languages. Zetawoof(&zeta;) 18:33, 29 June 2008 (UTC)


 * I just removed it again. —Piet Delport (talk) 2009-10-16 11:09


 * I would like to add in response to Hans Bauer that while lists in general may mention languages that do not deserve full articles, this particular list is for historically important languages, which should assume enough notability for an article. —Piet Delport (talk) 2009-10-16 11:20


 * Seed7 is listed as requested article for some time now. The article was deleted because of missing third-party coverage in 2006. Now Seed7 has third-party coverage. Citation from the article request: Paper from Daniel Zingaro: Modern Extensible Languages (alternate link) (paragraph about Seed7 at page 16), The Quest for the Ultimate Cycle explores the 3n+C extension of the Collatz Conjecture with Seed7 programs, A FreeBSD port / see also here (not maintained by Mr. Mertes). Additionally there is also: Abstract of diploma thesis (in german, about MASTER, a predecessor of Seed7), Doctorate thesis (in german, about MASTER, a predecessor of Seed7). And finally: The Seed7 Homepage. Maybe the new evidence is sufficient to undelete the Seed7 article. Hans Bauer (talk) 19:17, 16 November 2009 (UTC)


 * That's good to hear! (Just note that the article's recreation does not mean the entry here should be restored; see the proposed criteria below.) —Piet Delport (talk) 2009-11-18 02:31


 * There are many languages in this list, which are so new that they can hardly be classified as historically important. But I still consider all this languages important and they should be listed. Time flows by and Seed7 now has an article and it is described in a book with: "Seed7 supports the introduction of new syntax and their semantics into the language and it allows new language constructs to be defined using the Seed7 language itself" (reference: Abrial, Jean-Raymond and Glässer, Uwe, "Rigorous Methods for Software Construction and Analysis", ISBN 978-3-642-11446-5, Springer, 2010, page 166). Because of this syntactic and semantic extendability I consider it pioneering. The situation has changed and therefore I will add Seed7 back to the list. Hans Bauer (talk) 11:12, 12 July 2012 (UTC)

legend??
what do the 2 legends mean? especialy the first one. the 2rd one maybe means that "*" in 1st column means a unique language (no direct predecessor). It isnot necessary that " " appears here. in 1st legend, "(entry)" the bracket in the "name" column? who can help me? —Preceding unsigned comment added by Xujianguo (talk • contribs) 04:02, 27 May 2009 (UTC)
 * I agree, I don't see the logic behind the legends. Mahjongg (talk) 16:10, 11 June 2009 (UTC)

Definition of historic importance
In order to make this article more useful, and apropos the above discussion, i'd like to propose some objective inclusion guidelines. Comments and improvements are most welcome; it would be great if we could settle on a consensus version of this to put at the top of this talk page or somewhere.


 * Criteria
 * At minimum, entries should have an article that explains the language's historic importance. (Stubs or redlinks are acceptable as long as there is no doubt that an appropriate article could be written.)
 * In addition, entries should be at least one of:
 * Pioneering
 * Languages that innovated significant ideas, concepts, or paradigms (even if they were not otherwise influential or widely used).
 * The status of "significant" is somewhat subjective, but it should be reasonably widely accepted as such by neutral third parties: it's not enough for a language's authors to think it significant.
 * Examples: Plankalkül, ALGOL, IPL, Lisp, Forth, Simula, Prolog
 * Influential
 * Languages that inspired other historically important languages (even if they were not otherwise pioneering or widely used).
 * A minor language being the predecessor of several other minor languages is probably not important unless one of those descendants become historically important.
 * Examples: CPL/BCPL/B, ISWIM, Scheme, ABC
 * Widely used
 * Languages that found particularly widespread usage (even if they were not otherwise pioneering or influential).
 * One rule of thumb is any "x language" that's also an "x industry".
 * Examples: COBOL, C++, Java, C#, PHP, JavaScript

—Piet Delport (talk) 2009-10-20 08:28

Following the above, i'd like to propose a new column to the list that very briefly summarizes why the language is on the list, under a heading of "Significance" or similar. A sentence fragment from the language's article should suffice; for example, Plankalkül was the "First high-level non-von Neumann programming language". —Piet Delport (talk) 2009-10-23 00:44

Red link exclusion has become a rather serious problem. There are many otherwise well meaning editors who bulk remove "red links" from articles with the notion of "if it doesn't yet have an article it isn't notable". (I can cite specifics but it might offend a few people.) This is a logical fallacy in that often red links simply signify things that people have not yet written about. In the case of this specific article, the current red links all seem to be topics that really should be written about but are mostly older or historic languages that aren't getting too much attention right now. As you mention above, significance is subjective; I'd actually consider it to be highly subjective. This isn't a "timeline of historic programming languages" so trying to artificially limit inclusion based on pioneering, influential, and widely used doesn't sound like a good idea to me. Perhaps a better way (assuming this article has actually had a problem with inclusion, although I personally don't see a problem here) would be to simply limit inclusion to programming languages that have preexisting articles or would otherwise meet the notability guideline appropriate to this genre of articles should an article about the language be written? On a strictly editorial note, is there a reason to have so much bold text in this article? I find it seems to distract from the rest of the content. It also seems like the Predecessor(s) column should be the rightmost column and not the leftmost column. --Tothwolf (talk) 09:46, 23 October 2009 (UTC)
 * Having seen similar criteria implemented for other articles in the past, and having seen disastrous results, I'm going to offer my $0.02 here.


 * Hmm? This is exactly a "timeline of historically important programming languages", as the first line of this article states (and its content confirms, and the discussion above elaborates): its very usefulness depends on it. I'm not trying to change the selection criteria, i'm just describing it (as the stand-alone list guideline recommends).
 * Regarding "red link exclusion", nobody has suggested it: the proposal above specifically points out that redlinks are okay. (If you're concerned about the precedent of Seed7 above, the entry was removed because it was non-notable, not because it was a redlink.) —Piet Delport (talk) 2009-10-23 16:56

The article is named Timeline of programming languages and the lead section used to say "This is a chronological list of programming languages." and has been reworded numerous times up to the point where it is now. People are going to continue to add entries with the current lead. Something like what was in the lead in 2005 (except using a hatnote) might limit inclusion more than the current lead. Truth is, there is little, if any reason to limit a properly structured timeline article to only "historical" programming languages. Heck, I just now noticed how syntactically malformed the current wikitable code and section links are... --Tothwolf (talk) 18:04, 23 October 2009 (UTC)
 * No, I wasn't thinking of Seed7. I've seen a number of articles completely gutted of red links by editors using the reasoning I mentioned above.

Concept versus implementation
At the moment, a number of languages have separate entries, for "concept" and "implementation": "Fortran, IPL, Lisp, COMIT, COBOL, TRAC, APL, Simula, ALGOL, PL/I, Forth, Icon, Self"

While this is not necessarily a bad idea, i think it ends up cluttering the already-sprawling list without adding much extra value.

Proposal: Merge the "(implementation)" entries into their corresponding "(concept)" entry (excluding Plankalkül and ISWIM, which should keep their qualifier), extending their origin year range as necessary.

—Piet Delport (talk) 2009-10-20 17:37

Versions and variants
In addition to the concept versus implementation entries (see above), some languages have multiple entries for different versions or variants of them, which is probably overkill: it should be enough to list only the first or most prominent member of a closely associated line or group (unless more than one member can be argued to be separately historically important). Here's a quick, crude listing of entries that link to the same article, or otherwise appear closely related:


 * ALGOL: ALGOL 58, ALGOL 60, ALGOL 68, ALGOL W
 * Autocode: Mark I Autocode
 * BASIC: Altair BASIC, CBASIC, GFA BASIC, GW-BASIC, BASIC FOUR, PowerBASIC, QuickBASIC, STOS BASIC, True BASIC, Turbo Basic
 * C++: C with classes, Standard C++
 * COWSEL: POP-1
 * Common Lisp: ANSI Common Lisp
 * Distributed Application Specification Language: DASL (AUS), DASL (BOS)
 * ECMAScript: LiveScript, JavaScript
 * Erlang: Open Source Erlang
 * Forth: ANS FORTH
 * Fortran: FORTRAN "I", FORTRAN II, FORTRAN IV, FORTRAN 66
 * IPL: IPL II, IPL V
 * JOSS: JOSS I, JOSS II
 * MUMPS: Standard MUMPS
 * Modula: Modula-2, Modula-3
 * Oberon: Object Oberon, Oberon-2, Oberon-07
 * Pascal: Object Pascal, Turbo Pascal, Embarcadero Delphi, Component Pascal
 * REXX: NetRexx, Object REXX
 * SNOBOL: SNOBOL3, SNOBOL4
 * Smalltalk: Smalltalk-76, Squeak
 * Speakeasy: Speakeasy-2, Speakeasy-3, Speakeasy-IV
 * ML (programming language): Standard ML, F Sharp (programming language)
 * Visual Basic: Visual Basic .NET

Many of these can probably be merged; any comments on which ones shouldn't be?

—Piet Delport (talk) 2009-10-20 22:04


 * Don't be led astray by similar naming: Visual BASIC.NET has far more in common with C# than with Visual BASIC, so it would be better merged with the former than the latter. Algol60 and Algol68 are actually pretty different as well. Also the BASIC article was split from the various implementations on the basis that it was too large. It would not be a good idea to merge them all back together. -- Derek Ross | Talk'' 00:57, 23 October 2009 (UTC)


 * I agree, don't be led astray by names. e.g. Component Pascal is actually a superset of Oberon-2; merging Modula-2 and Modula-3, or (less so) Oberon with Oberon-2 would make as much sense as merging C with C++ etc. Chris Burrows (talk) 05:39, 23 October 2009 (UTC)


 * Right. I did not suggest merging all of these: it's just a semi-automatic list of related entries as a starting and reference point. It's as important to know why some similar-looking entries warrant separate inclusion as it is to know why some do not. (In case there was confusion, i definitely did not suggest merging any articles, just entries here.)


 * Regarding the incarnations of BASIC, i was not talking about merging the articles themselves (which is obviously a bad idea), but about removing entries from this list, which is definitely not the place for every dialect and implementation. (Interested users can find those in the BASIC article, not here.)
 * Looking through the list, the candidates with most historic significance seem to be:
 * BASIC itself, of course (Dartmouth BASIC)
 * Altair BASIC, for being Microsoft's first product, and the subject of Bill Gates' historic Open Letter to Hobbyists
 * The others can probably go, unless one of them also has clear and independent historic significance. (Any suggestions?)


 * Regarding VB and VB.NET, the differing implementations do not matter much; they are still essentially releases of the same language, with the same (shared) historic significance. An entry for the genesis of Visual Basic should be sufficient; interested users can find all the different implementations and releases of the language through the article.


 * Thanks for the note about Component Pascal's lineage. I can't tell from the article what kind of historical significance it has, though; do you know?
 * It is an extension of Oberon-02 primarily suited to the implementation of Component Frameworks - hence its name. It was one of the original seven languages selected by Microsoft for implementation on its .NET framework.Chris Burrows (talk) 00:53, 24 October 2009 (UTC)
 * The Modulas seem to be well-motivated:
 * Modula-3 for its influence in research circles and in the designs of languages like Python, Java, and C#
 * Modula-2 for its influence in Modula-3, the Oberon family, and other languages
 * Modula for being the forerunner of the above
 * I'm also not sure how to describe the Oberon family, but they seem to continue the line of Wirthian innovation; it would be great if someone more familiar with them could give a summary of what makes each of them significant. (Thanks for the response so far!) —Piet Delport (talk) 2009-10-23 15:45


 * Within a couple of years of the appearance of Oberon the additional O-O concepts researched in Object Oberon were added to it to form Oberon-2. It is suitable for general-purpose systems and application development. Oberon-07 also evolved directly from Oberon via Oberon-SA (StrongARM) with features that make it particularly suited to the development of reliable software on resource-limited systems e.g. real-time embedded control systems using microcontrollers.Chris Burrows (talk) 00:53, 24 October 2009 (UTC)


 * Thanks for the clarification. That seems reasonable. -- Derek Ross | Talk'' 15:56, 23 October 2009 (UTC)


 * Altair BASIC is certainly historical, but the criterion for deciding whether something is a dialect or a distinct language isn't based on historicity. Altair BASIC looks like a very basic subset of Dartmouth BASIC. The BASIC entries in the list seem to over-represent Microsoft as well. Why not Coleco ADAM BASIC, or a hundred others? MSX BASIC doesn't seem like a distinct flavor either, it's a dialect/subset. Also, why is there no mention of Visual FORTRAN? Or Visual COBOL (or did I miss it?)? The "visuals" represent a specific historic stage in the evolution of prgramming following the almost universal adoption of GUI operating systems, don't they? Almost all of the comments on the Talk page seem well considered and considerate and I hope mine don't deviate, but I do detect, rightly or wrongly, a little bit of bias at work in the published product.Hypatea (talk) 19:08, 30 May 2010 (UTC)

C-10 nomenclature
Betty Holberton, "Holberton worked with John Mauchly to develop the C-10 instruction for BINAC ...". See the second clause which reads as a run-on sentence and a major claim. But if you read her interview for an oral history, she never claimed to have invented a programming language. Rather, it is an instruction set. Read the 1948 Research Report " A Logical Coding System Applied to the ENIAC " for many other kinds of instructions, such as C-1, C-5 etc.

Holberton does claim C10 was an assembler instruction set for Univac I (bought out BINAC) which starts with mnemonic letters S=subtract, T=test. --Ancheta Wis (talk) 07:09, 28 August 2010 (UTC)

Based on Holberton's interview, it appears that this is the instruction set for a UNIVAC. --Ancheta Wis (talk) 07:45, 28 August 2010 (UTC)

OpenCL and CUDA for programming on SIMD processors architecture
OpenCL and CUDA just appeared recently. They are, probably, the first commercial success among programming languages for SIMD architectures. I think they might be in the list as well. —Preceding unsigned comment added by 82.8.176.196 (talk) 19:04, 18 September 2010 (UTC)


 * As far as i understand, OpenCL is more of a programming framework / hardware architecture abstraction layer (analogous to OpenMP) than a programming language proper, and probably doesn't belong here. --Piet Delport (talk) 09:59, 7 January 2011 (UTC)