[antlr-interest] what would ANTLR ref guide revised edition have?

[bill punch]

Let me say that I really like ANTLR and have been using both books
extensively. Thus I hope this is taken as constructive criticism.

I think what Ray is saying is what I've been running into. I'm teaching
the Compiler's class (for the first time) and using the ANTLR reference
book. It has been a bit of a struggle to get things in an order that I
think they can easily assimilate. I want to focus on the higher-level
issues and less on the details of the underlying programming, until it
has to be addressed. For example, the progression so far has been :
lexer/parser for a simple language, output an AST for that simple
language, now interpret (AST->Java Tree Walker) with global scope, no
functions. Next I hope is language with scope, then probably translation
to (probably) Java bytecode using templates. Each project adds a little
more of the language, so we have something at the end. I'm working with
Logo at the moment, with some Java/Swing for the turtle.

What I've found is that I need something in-between the two books. The
reference book has most of the stuff I need, but a bit spread around and
with what seems like a lot of detail in between. The LIP has far too
much detail in many ways, which makes getting the point across difficult.

As with Ray, I would prefer a better progression like the LIP book, but
at a level above the Java details that come up. One can go quite a ways
without having to write a lot of Java, and then it can be introduced
slowly. They didn't really have to write much Java till the interpreter,
and even then I took the Pie example and stripped out most of the scope
stuff (took out functions, error checking) to get down to a basic
skeleton. I think a lot of that material could have been discussed
(Symbol, Scope, resolve, etc.) at a higher level and then shown
implemented in Java. I think a lot of the details get caught up in the
language and it really shouldn't have to.

If done as described, I suppose it would be easier to get other targets
in, though I think if the concepts are clear and one implementation is
provided, others would be easily done (perhaps as appendicies).

Anyway, it has been something I've been thinking about.

      >>>bill<<<


On 2/15/11 8:40 PM, Terence Parr wrote:
> Hi Ray,
>
> Yeah, the language implementation patterns book is much more amenable to a class I would say. the reference guide is more of an accompanying text. do think that is working out for you this semester would be examples in the LIP book?
>
>  I'm actually teaching a programming language is classed myself this semester. I have been doing six projects that take anywhere from 3 to 10 hours depending on their abilities. This gives them the skills necessary to tackle the Smalltalk compiler and interpreter starting next month. yikes. I hope it does anyway ;)
>
> Ter
> On Feb 15, 2011, at 5:01 PM, Raymond J. Schneider wrote:
>
>> An orderly sequence of structured examples of increasing difficulty illustrating the concepts covering all the bases would be nice.  This is especially true as one enters the realm of trees and then string rewriting.  I used the book for a class on writing language parsers and we had a good deal of difficulty after the first part.  We're currently using the Language Implementation Patterns book with a bit more success so far because of the orderly presentation that the use of patterns is providing.
>>
>> Regards, Ray
>> --
>> Ray Schneider,PE, Ph.D
>> Associate Professor
>> Math and Computer Science
>> Bridgewater College
>> http://www.bridgewater.edu/~rschneid/
>> http://theweedlessgarden.blogspot.com
>> ________________________________________
>> From: antlr-interest-bounces at antlr.org [antlr-interest-bounces at antlr.org] On Behalf Of Terence Parr [parrt at cs.usfca.edu]
>> Sent: Tuesday, February 15, 2011 18:36
>> To: antlr-interest Interest
>> Subject: [antlr-interest] what would ANTLR ref guide revised edition have?
>>
>> Howdy. I'm thinking about revising the ref guide for ANTLR 3 (ANTLR v4 might be awhile so I should update book).  Any suggestions to improve?  One obvious thing: discuss not just java target :)
>>
>> Ter
>>
>> List: http://www.antlr.org/mailman/listinfo/antlr-interest
>> Unsubscribe: http://www.antlr.org/mailman/options/antlr-interest/your-email-address
>
> List: http://www.antlr.org/mailman/listinfo/antlr-interest
> Unsubscribe: http://www.antlr.org/mailman/options/antlr-interest/your-email-address