How realistic are the examples in Language Implementation Patterns

[Ashok C. Mohan]

I read from amazon that the implementation patterns are in Java. How realistic is this? Are there any programming languages around which are implemented in java? Or did I get the whole implementation pattern confused with the implementation of the language itself? What is the most widely used languages for implementing languages? Will this book get me started in writing a PL for robotics (for example)? Thanks and Welcome to Terence Parr. Welcome to javaranch.

[Ulf Dittmer]

Note that "language" is not the same as "programming language". The book deals specifically with domain specific languages, many of which have been implemented in Java (and in all kinds of other programming languages). If you visit the web sites of the major Java parser generators (Antlr, JavaCC, SableCC) you'll see the range of grammars for which parsers have been built.

[Burk Hufnagel]

Ashok,
I may be wrong, but I believe much of Groovy is written in Java.
Burk

[Terence Parr]

Ashok C. Mohan wrote:I read from amazon that the implementation patterns are in Java. How realistic is this? Are there any programming languages around which are implemented in java? Or did I get the whole implementation pattern confused with the implementation of the language itself? What is the most widely used languages for implementing languages? Will this book get me started in writing a PL for robotics (for example)?

Hi Ashok, the book is about implementing new DSLs and general purpose programming languages. So, for example, to implement a PL for robotics (which I did in the 80s and was used by Renault to build cars), you could implement the PL in any language including assembly language. I had to choose an implementation language for the book so that I didn't flip back and forth between a number of implementation languages.

The choice of implementation language often depends a great deal on how efficient the language application has to be. For example, I built a byte code compiler and a byte code interpreter in C for the robot control language. Using C let me make a fairly efficient interpreter, but I could've used [shudder] Fortran to write the compiler itself.

[Burk Hufnagel]

Terrance,

I don't think it's what he meant by realistic, but I'm wondering if the examples are all very simple (so that we can understand them easily) or if there are a few complex enough that they could be used for real-world (read business) solutions.

Thanks,
Burk

[Terence Parr]

Java is a realistic language used is implementation mechanism. The examples are not trivial but are not as big as one might want an industry. Their medium-size is a good compromise between learning and completeness.

[Burk Hufnagel]

Terence Parr wrote:Java is a realistic language used is implementation mechanism. The examples are not trivial but are not as big as one might want an industry. Their medium-size is a good compromise between learning and completeness.

So they're complex enough to prove they can do some "heavy lifting" while being simple enough for a beginner to grok?

Do they include error handling (how to deal with improper syntax, or other errors) or do they assume correct input and leave the implementing the error handling as an exercise for the reader?

Burk

[Terence Parr]

Burk Hufnagel wrote:Do they include error handling (how to deal with improper syntax, or other errors) or do they assume correct input and leave the implementing the error handling as an exercise for the reader?Burk

It talks about error handling in the parsing patterns but the (ANTLR reference talks more about this as error handling is often highly specific to a tool. It DOES, however, talk about how to compute type information and detect operations with incompatible types. Semantic errors not syntax errors.

[Burk Hufnagel]

Terence Parr wrote:It talks about error handling in the parsing patterns but the (ANTLR reference talks more about this as error handling is often highly specific to a tool. It DOES, however, talk about how to compute type information and detect operations with incompatible types. Semantic errors not syntax errors.

That's good, it always bothers me then I can't find any samples of how to deal with error conditions when I'm learning something new. Makes me wonder if the author was afraid that showing what it looks like in the real world would scare off the reader. What's worse, many developers will use the example as a pattern for how they should implement something and not include any error handling capability at all.

Thanks, Terrence.

Burk