An inclusive name for complete units of code?

[Steven T. Hatton]

Is there a formal name used to denote every syntactically complete part of a 
program?  I'm trying to describe something I'm calling a 
"translation-subunit".  What I mean by that is a grouping of code that any 
reasonably sane person would make the complete contents of a source file.  
IOW, I'm not talking about something along the lines of

//------------------------
//start.h
namespace fubar {
  //fubar definitions
// END OF FILE


//-----------------------
//end.h
   //assume fubar bracket is still open
  //more fubar definitions
} // end of fubar namespace
// END OF FILE

//------------------------
//fubar.h
#include "start.h"
#include "end.h"
// END OF FILE


But I am talking about:
//------------------------
//subunit.h
namespace unit {
  //unit definitions
} // end of fubar namespace

// END OF FILE

When I tried to express what I meant by this notion of a translation-subunit, 
I found myself lacking a word to use to denote what should go in it.  I'm 
trying to abstract the concept of a (typical) source file or header file 
which and talk about it in terms of how its content fits into the syntactic 
unit called a translation unit.

As for the word I'm looking for, I'm /mostly/ thinking in terms of stuff 
enclosed in bracketing constructs such as parentheses, curly braces, square 
brackets, angle brackets, etc. and usually preceeded by a declarator id. (If 
I'm using that term correctly.) I'm pretty sure the general idea would be 
synonymous with the set of all nodes in the parse tree of a grammatically 
correct program.

-- 
Regards,
Steven