[antlr-interest] TreeDL way

[Andy Tripp]

Alexey Demakov wrote:

> Hi all,
>
> My project, TreeDL (http://treedl.sf.net), is mentioned in the article 
> "Manual Tree Walking Is Better Than Tree Grammars" by Andy Tripp. I 
> generally agree with this statement
> but can't agree with the context:
>
>> The mapping from one AST structure to another is quite complex
>> (see XSLT, TXL, or TreeDL for example), and a single BNF-like grammar 
>> doesn't capture it.
>
>
Hi Alexey,
Yea, I guess I shouldn't mentioned TreeDL. Thanks for correcting that - 
I'll take it out.
As you can see, I was just trying to list a few technologies relating to 
tree-transformation.

As for the rest of your comments, I just have one place where I'm very 
sceptical:

>
> 2a. It is source of code generation. You need not write depth-first
> walker by hands and update it on each AST change - just re-generate it.

I'm not sure exactly what you're claiming here. You're not saying that, 
for example,
I could change ANTLR to produce a different AST structure for Java code 
input,
and have whatever AST-processing I've written still work correctly, are you?

I mean, if the AST structure changes, whatever code is written that does 
something
with that AST (whether just print it, change it, or whatever) is now broken.

Andy