Ok, a few out there know that I’ve actually been working on my ADAM language (Advanced Data Acquisition and Manipulation) for some time. I had actually had a lexical analyzer, byte code compiler, and base interpreter already built. The only thing I didn’t like was the definition of the language. It was leaning toward an imperative language, and this is not exactly the vision I had for it.

So ADAM is going for a complete reconstruction. Not only is the language definition closer to what I was envisioning, but I am now moving toward a .NET compiled language. This will be accomplished by changing the byte code compiler to compile to MSIL instead. MSIL, for those who do not know, is the Microsoft Intermediate Language. This is what all .NET language compile to before being compiled to an executable. MSIL resembles C++ in that the syntax is very similar, complete with classes, and the order of defining methods and variables. Other parts are similar to assembly language in that you must push all your function parameters prior to calling, or JUMPING in assembly, your function call.

Ok, so I’ve got a better example of the language structure, as well as a sample breakdown of the lexical analysis output tree.

create source Employees  
begin  
"excel" -\> type  
"employees1099.xls" -\> filename  
"Sheet1" -\> sheet  
"FirstName != ''" -\> criteria  
end

COMMAND : CREATE SOURCE  
PARAMETER : Employees  
IDENTIFIER : BEGIN  
STRING : "excel"  
DELIMITER : -\> (assignment)  
IDENTIFIER : type  
STRING : "employees1099.xls"  
DELIMITER : -\> (assignment)  
IDENTIFIER : filename  
STRING : "Sheet1"  
DELIMITER : -\> (assignment)  
IDENTIFIER : sheet  
STRING : "FirstName != ''"  
DELIMITER : -\> (assignment)  
IDENTIFIER : criteria  
IDENTIFIER : END

This would open up an excel sheet as a data source. The objective here is to make the language as “descriptive” as possible. Stay tuned for more work on this.

I now have to spend a bit more time defining the token representation of the language, the different elements it will break down into, and then the structures that will house this data. Then I must map out how to move this to MSIL efficiently.