Language Definition vs. Code File Output

Hi Nassim

1) In terms of a dynamic lexer, the code-based one will just be slightly faster to load than from XML since parsing XML takes longer. But both will contain the same lexing data, classification tagging capabilities, and example text provider.

2) For code generation, the Language Designer helps in these areas:

Dynamic lexer (full implementation)
Programmatic lexer (ID classes only)
Token tagger
Classification type provider
Example text provider
Syntax language stub class that can load all those things

The Language Designer can also generate type-specific AST node classes. If you will be using our LL(*) Parser Framework, we recommend you generate the AST node classes with the Language Designer since it only takes a few things as input and creates complicated AST class files that work perfectly with our parser framework.