May POSTPONE cause arbitrary parsing?

[ruv]

The compilation semantics for POSTPONE, an excerpt:

Skip leading space delimiters. Parse name delimited by a space. Find name. Append the compilation semantics of name to the current definition.

If the Recognizers word set is provided, it can be something like the following:

• Skip leading space delimiters. Parse lexeme delimited by a space. Recognize lexeme. Append the compilation semantics of lexeme to the current definition.

If a recognizer does additional parsing, it changes the lexeme, i.e., the produced token will determine compilation semantics for the different lexeme.

So, there is no a place for additional parsing. And it should be forbidden for a recognizer to do additional parsing.

[ruv]

Anton's response[2] to above resoning[1] was:

Your "recognize lexeme" invokes the string recognizer, which parses ' bar"'

I.e., that "to recognize" means invoking a recognizer, and it doesn't matter what the recognizer does, since it may do anything in terms of side effects (including additional parsing).

It looks like circular reasoning in terminology:

• a. "To recognize" is to perform a recognizer.
• b. A "recognizer" is a definition that is performed to recognize.

My objection[3] was that Anton used the terms (that I defined) in another meaning to confute my reasoning. It's incorrect.

You have to either use my terms, or define your terms and show that your model is better.

[1] news:[email protected]
[2] news:[email protected]
[3] news:[email protected]