From: bc@freeuk.com   
      
   On 29/10/2025 15:36, David Brown wrote:   
   > On 28/10/2025 20:14, Janis Papanagnou wrote:   
      
   >> Reasons for dissemination of a language are multifold; back then   
   >> (but to a degree also today) they were often determined by political   
   >> and marketing factors... (you can read about that in various historic   
   >> documents and also in later ruminations about computing history)   
   >   
   > I can certainly agree that some languages, including Algol, Algol 68 and   
   > Simula, have had very significant influence on the programming world and   
   > other programming languages, despite limited usage.  I was interpreting   
   > "useful programming language" as meaning "a language useful for writing   
   > programs" - and neither Algol 68 nor Simula are sensible choices for   
   > writing code today.  Neither of them were ever appropriate choices for   
   > many programming tasks (Algol and its derivatives was used a lot more   
   > than Algol 68).  The lack of significant usage of these languages beyond   
   > a few niche cases is evidence (but not proof) that they were never   
   > particularly useful as programming languages.   
      
   Algol68, while refreshingly different when I came across it in the late   
   70s, was a complex language.   
      
   Its reference document, the Revised Report, its two-level van   
   Wijngaarden grammar, suggested a language too much up its own arse.   
      
   Its complexities tended to leak even into straightforward features that   
   people are familiar with from other languages.   
      
   Understanding it, and confidently using it, looked hard. Implementing it   
   must have been a lot harder.   
      
   Also, at the time I'd only ever seen examples of it in print, where it   
   was beautifully typeset and looked gorgeous.   
      
   The reality when I finally got to try it was very different. You spent   
   half the time fighting with upper/lower case and trying to get   
   semicolons right. And most of rest grappling with esoteric error   
   messages couched in terms from the revised report (which has its own   
   vocabulary).   
      
   I borrowed some syntactic features I considered cool, but I had to   
   produce a real, practical systems language for microprocessors, whose   
   compiler had to run on the same machine.   
      
    From this perspective, I consider it rather dreadful now, with lots of   
   dubious-sounding aspects.   
      
   Take this one: comments start with '#' (an alternative to COMMENT) and   
   also end with '#'. Leave out '#' (or have a stray one) and everything   
   now gets out of step.   
      
   Or this one:   
      
      print((2 + 3 * 4));   
      
      BEGIN   
         PRIO * = 5;   
         print((2 + 3 * 4))   
      END   
      
   The first print shows 14. The second shows 20, as the precedence of '*'   
   has been set to match that of '+'.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)