DDA Computing elements

The DDA language is built around the analog computing elements (or primitives). These are basically electrical block circuits implementing basic arithmetics such as summation and multiplication, but also integration. These are also special elements for clipping, exponentials, square roots, and many more.

The DDA domain specific language is agnostic for function names. Host languages such as Python or C are not. Many of the function names are reserved words in these languages. Examples are:

  • const: Reserved word in C/C++ for constant variables

  • int: Type name in C, overwritable in python

  • sum: Builtin in python

  • div: Function in stdlib.h in C, which can cause clashes in slightly more complex C codes.

When it comes to exporting to languages such as C, we rewrite these keywords. In Python, we don’t have to, because none of the well known DDA function names is really reserved. The primitive builtins can always be recovered by from builtins import int, sum, etc.

Different ways to access the well-known DDA computing elements in Python

If you write

>>> from dda.computing_elements import *
>>> int(int,sum)  # Make use of the imported Symbols
int(int, sum)

you will load a bunch of names such as int and sum in your local namespace.

You can also just call

>>> from dda.computing_elements import dda_functions, dda_symbols
>>> print(dda_symbols["floor"]) # will print a Symbol()

You can use the symbols dictionary to populate your namespace at whish:

>>> globals().update(dda_symbols)

Note that Python globals are module-local, so we cannot provide this line as a function.

Last but not least, you can also just use the namespaced version with prefixes, which leaves you on a safe footing:

>>> dda.floor(dda.sum)

Definition/Implementation of the primitives

Once we have a pure-Python DDA evolution code (probably using scipy), we will have a python implementation of the DDA functions. Otherwise I could avoid that. This module also contains a C++ implementation of the primitives, which resides as a string (cpp_impl).