Parser example in SableCC, ANTLR, Parsec

Scenario 1: Instantiate user-defined datatypes

--> SableCC fails. No problem in ANTLR, Parsec.

Scenario 2: Store scale factor as integer

--> SableCC fails. No problem in ANTLR, Parsec.

Scenario 3: Abstract over commonalities in graph productions

--> SableCC fails and ANTLR fail. No problem in Parsec.

Scenario 4: Reasoning about static/dynamic errors in code snippets

--> abstraction barrier breaks in ANTLR


Regions example: Interpreter, embedding, PE

Scenario 1: Language extension 

--> easy with embedding, but possible in all approaches

Scenario 2: Language composition

--> easy with embedding and PE, not obvious how to do it with interpreters

Scenario 3: Add alternative interpretations

--> easy with interpreter and PE, not possible with embedding

Scenario 4: Domain-specific optimizations

--> easy and flexible with interpreter,
    possible via alternative interpretation with PE
    nonmodular and complicated with embedding

There are more differences when one considers names, embedding type systems etc.