ℙ𝕖𝕡 🙴 ℕ𝕠𝕞

Writing a blog changed my life. Jeff Atwood, somewhere on his blog

the current tape cell in the ℙ𝕖𝕡 machine

About the current tape-cell in the tape array

I refer to the array of attributes that correspond to each parse-token on the parse stack as a tape . I think this is non-standard terminology. By a 'tape' I mean an (possibly infinite) array of mutable strings that includes a current valid pointer to one of the array cells.

The tape-cell that is currently pointed to by the tape-pointer I refer to as the “tapecell” or “tape-cell” or “current cell” etc. The current cell of the tape is decremented by the pop command, incremented by the push command and also can be incremented and decremented with the ++ and -- commands.

The pop and push commands are designed so that the current tape cell is unchanged when a pop is called on an empty parse stack, and also unchanged when a push is called on an empty workspace buffer.

This behaviour is designed to keep the pep machine stack and the tape “synchronised”. That means that the tape-pointer should point to the last parse-token popped into the workspace buffer.

   # nom fragment
   pop;pop;pop;
   "size*angle*end*" {
     # The current tape cell points to the attribute 
     # corresponding to the size* parse token
     # do something with the attributes or parse tokens
     nop;
   }
 

The ++ and -- commands are used to “compile” or translate the parse token attributes when a shift-reduce rule is applied in a block