Wikipedia:Reference desk/Archives/Computing/2019 January 14

= January 14 =

Popularity of LISP in 1989
If you look at TIOBE, scroll down to "very long term history", it says that LISP was the second-most popular programming language in 1989 (and remained quite popular for more than 10 years after that). I can understand why COBOL was #3 in 1999, all of the code that had to be changed for the Y2K problem, but I don't see why LISP was so popular in 1989. Does anyone know? Bubba73 You talkin' to me? 19:39, 14 January 2019 (UTC)
 * Those are quite dubious listings, as they don't explain their methodology, or list all the languages studied. Visual Basic was huge through the 1990s, but they only (in that summary) show the .NET version. BASIC (generically) was still very popular through the '80s. Where's Modula-2? dBase II? Why isn't SQL there long before 1999? They also show Pascal (generically) climbing from 1989 to 1994, yet dropping out by 2000. I presume that's mostly Object Pascal, but Turbo Pascal went much the opposite way - increasingly popular under MS-DOS, but then never recovering its pre-eminence once Windows 3.0+ shipped after 1990.
 * As to Lisp, then Lisp (and Smalltalk) was talked about a lot in the late '80s, but most people were using 286 PC clones or 68000 workstations if they were lucky, never had enough RAM, and didn't have the performance considered to make it worthwhile. Apart from US universities (MIT being a leader, although mostly with Scheme), one platform that did use a lot of Lisp invisibly was AutoLISP embedded within Autodesk products such as AutoCAD.
 * I'd be looking at old copies of Byte for a more detailed view of this. Andy Dingley (talk) 20:31, 14 January 2019 (UTC)


 * Yes, that Pascal ranking is strange #14, then #3, then #5, then #96, then #14 twice, then #199. Perhaps they split off Delphi and Object Pascal.   used to have a lot of Byte magazines, but I got rid of them a few years ago.  I used LISP in an AI in the 1980s, and it did take a lot of memory. None of us could run our programs until the professor got our memory limit on the mainframe increased. Bubba73 You talkin' to me? 21:48, 14 January 2019 (UTC)


 * I find the listings at the TIOBE website to be entirely useless, because they implicitly assume that any language is suitable for any task.
 * For example: the "percentage" of the marketplace who program in VHDL cannot be meaningfully compared the percentage who program in Objective-C or in the "R" programming language. These communities are quite different in demographic, in purpose, and in the type of technical work that the programmer is doing.  If somebody recommends that you switch from "R" to "VHDL" on the basis of language-popularity, you should stop taking technical advice from them!
 * Language "popularity" is a very silly metric; computer languages are not generally "interchangeable" in the context of a practical setting.
 * A much more useful way to approach this topic is to read about Lisp (and its history) from the perspective of an expert: for example, Peter Norvig's Paradigms of AI Programming retrospective is very informed; John McCarthy's LISP---NOTES ON ITS PAST AND FUTURE is very well-informed; and so on.
 * These aren't just random junk-links found via an internet search; these guys (Peter Norvig and John McCarthy) are not just random bloggers or casual commentators - they happen to be important figures in the history of the LISP programming language. They have useful perspectives on the language, and its implementations, and the communities who used and modified them.  When a world-expert in optimization theory and formal language chooses to apply the words "approximate local optimum in the space of programming languages;" or, "In 1991 Lisp offered a combination of features that could not be found in any other language"... each word in these summaries carries great meaning.  These are direct quotes from people who studied (and invented) lots of computer languages.
 * Nimur (talk) 21:50, 14 January 2019 (UTC)


 * the 1980's were the time of Lisp machines and high enthusiasm about AI (before the 1987 AI winter.) Aecho6Ee (talk) 13:55, 15 January 2019 (UTC)


 * Yes, but that article says that there were only about 7,000 Lisp machines. Bubba73 You talkin' to me? 18:41, 15 January 2019 (UTC)