Anonymous user
Category talk:Plain English: Difference between revisions
→Clutter?: A joke language without a working compiler??
(Response about clutter.) |
(→Clutter?: A joke language without a working compiler??) |
||
Line 9:
:Regarding your concerns about efficiency, I am somewhat mystified why this is even a concern for the (intentionally) trivial tasks I have been submitting to (for now). How much more efficient would you like basic language constructs, loops and type conversions to be? --[[User:Chunes|Chunes]] ([[User talk:Chunes|talk]]) 11:36, 16 September 2020 (UTC)
:: I have looked at the Plain English compiler written in Plain English, and it's nice, very nice indeed. But where would one find a working executable compiler? Without one, it's hardly possible to test that any of the task implementations actually work correctly.
:: I have found and tested two executables purporting to be a working Plain English compiler: <code>cal-3040.exe</code> and <code>cal-4700.exe</code>, found at [https://github.com/Folds/osmosian/] and [http://www.osmosian.com/cal-4700.zip] respectively.
:: The former appeared to work at first, seemingly being able to compile the compiler itself, producing an executable named <code>osmosian.exe</code> in the compiler folder, which works exactly as the original <code>cal-3040.exe</code>. In fact, it is ''exactly'' the same file, byte for byte. And that was a bit suspicious, so I tried the "official" sample program <code>source.txt</code> obtained from [http://osmosian.com/sample.zip]. And I was not very surprised when the result of compiling <code>source.txt</code> was again an executable named <code>osmosian.exe</code>; again a byte-for-byte copy of <code>cal-3040.exe</code>. In other words, the compiler didn't compile anything, it just created a copy of itself named <code>osmosian.exe</code>. Impressive.
:: Then I looked for another version of the compiler, thinking maybe <code>cal-3040.exe</code> is just buggy, and I found <code>cal-4700.exe</code>. The name suggests it is a newer, better version. Indeed, this one has a new feature – it only makes a copy of itself when ran on its own source. When compiling <code>source.txt</code>, it showed an error message <code>I need a routine to 'initialize before run'.</code> and besides that produced no output. Then I tested a few task implementations from RC. I picked those which were supposed to produce output, and thus prove that the program – and hence the compiler – works correctly. These tasks all contain a definition <code>To run:</code>, and they all ended up in an error message <code>Error in the desktop. I already know how to 'run'.</code> without producing any other output.
:: At this point I am having a hard time believing this language is not a joke, though I hope I am wrong because the idea is enticing and someone has obviously put extraordinary effort into the whole... thing. There's 120 pages of documentation and a grammar description in EBNF; source code of the compiler which is over 23,000 LoC (excluding empty lines and comments); there's the executable "compiler" which is more of an IDE, with user interface providing such functionality as copying files, creating new folders, sorting or reversing selected lines of a file and all kinds of stuff that one would not expect even from somewhat advanced text editors; there is a rather sizable blog [https://osmosianplainenglishprogramming.blog/]; and then there is the mentioned sample application, with source code, and an executable which actually works... (but was it compiled from the Plain English source or written in another language?) It seems there's pretty much everything, except a working compiler. And under these conditions, I consider Plain English '''clutter'''. --[[User:Dick de Bill|Dick de Bill]] ([[User talk:Dick de Bill|talk]]) 18:25, 20 September 2020 (UTC)
| |||