Pep and Nom

the ℕ𝕠𝕞 "delim" command

The delim command changes the nom parse token delimiter to the given character.

Any character can be used for the parse token delimiter including spaces punctuation or any other character. It is a matter of personal preference. I like '*' which is why is it the default. The pep virtual machine and the ℕ𝕠𝕞 language are based on the idea that

Everything is text mjb, the inventor of nom

Even zeros and ones are probably text if you are looking at them... But the significance of this is that parse tokens are also text, not some mysterious en.wikipedia.org/wiki/abstract_syntax_tree or some other fancy data structure that I (and probably you) don't really understand. But you understand text , correct? Well at least we think we do: it’s some kind of ascii or UTF8 or whatever encoding that is hopefully terribly Unicode in the nicest way.

In nom you just build parse tokens like you would any other piece of text and then push them onto the stack and leave them there until you need them again. Isn't that great? Its one of the strangely simple but revolutionary ideas of nom*. But I am digressing; lets get back to the delim command.

   delim ".";
   delim '.';  # no difference
 

The parse token delimiter governs how the push and pop commands work. If the delimiter is ',' and the workspace buffer is

 one,two,three,

 two,three,

However, if we change the delimiter to a “.” dot and then issue another push command, then the whole workspace will be pushed onto the stack

   # workspace is now "two,three,"
   delim "."; 
   push;
   # workspace is now "" (ie empty)
 

The default (starting) parse token delimiter is '*' and normally you would only change the delimiter once at the start of the nom script to suit your visual preference.

  begin { delim '#'; }
  read;
  # write your script here
  (eof) {
    # do something when the end of the input 
    # stream is reached
  }
 

There is only one practical exception to this rule-of-thumb that I know of: To use the delim command as a hackish way to “split” the workspace string.

   "two words" {
     delim " "; push; 
     # workspace is now 'words'
     add "\n"; print; clear; 
     delim "*"; pop;
     add "\n"; print; clear; 
   }
 

The Nom language doesn't have any particularly fancy commands to work on the workspace buffer (and this was a deliberate choice for the sake of speed and simplicity), so sometimes it is necessary to resort to hacks like the one above.

notes

The delimiter register combined with the pop and push commands is a powerful way to manipulate grammar parse-tokens within a text-based virtual machine. It means that the same engine (pep) that is used to manipulate the token attributes is also used to manipulate the parse tokens, which results in a fast, simple and transparent parsing system

see also

• push push a token onto the stack
• pop pop a token. you can do it.
• /doc/commands/ docs for all the commands