[antlr-interest] "Language design patterns" book available in beta
Terence Parr
parrt at cs.usfca.edu
Fri Jun 5 16:01:38 PDT 2009
I guess it all comes down to how we defined simple DSL ;)
Yeah, stuff like configuration files are pretty easy. The interpreters
I have in the high-level interpreter chapter are suitable more for
dynamically typed DSLs because they are not super speedy. They just
happen to have the fewest moving parts. The bytecode interpreter
chapter is better for larger programming languages; more moving parts
because it needs a source code to bytecodes translator and then the
interpreter itself.
Ter
On Jun 5, 2009, at 3:36 PM, Randall R Schulz wrote:
> On Friday June 5 2009, Terence Parr wrote:
>> On Jun 5, 2009, at 2:29 PM, Steve Cooper wrote:
>>> First, thanks for writing this book. I'm a good way through already
>>> and have written my first LL(2) recursive descent parser today. I'm
>>> very happy with the book.
>>
>> great!
>>
>>> "If I have to skimp to make it into only 350 pages, where should I
>>> be terse?"
>>>
>>> I can only offer this; most people, I would think, will be writing
>>> smaller languages and DSLs. So deemphasise anything that applies
>>> only to large general-purpose languages.
>>
>> true...
>
> Is it?
>
> Simple languages (such as typical DSLs) are simple to parse and
> process.
> I would think there's more to say about more complex languages with
> richer grammars that lead to more demanding language processing issues
> in areas such as error-handling and -reporting, ambiguity resolution,
> symbol table design, island grammars, AST design, etc.
>
> So I would say aim towards the more complex, less obvious stuff.
> TDAR is
> sufficient for dealing with simple languages, isn't it?
>
>
>> ...
>>
>> Ter
>
>
> Randall Schulz
>
> List: http://www.antlr.org/mailman/listinfo/antlr-interest
> Unsubscribe: http://www.antlr.org/mailman/options/antlr-interest/your-email-address
More information about the antlr-interest
mailing list