4.5.1 Current C Interface

The current C interface is admittedly not the best, not the cleanest and not the most straightforward possible.

While the libaime interface addresses the concerns of its client applications and is reasonably well articulated, it does exposes a lot of subsystems and protocols and it requires them to be observed. Integrating the aime interpreter into an application is tedious.

Typical examples include the requirement for the application to define a variable set (be it even void) when no variables are defined. Or to indicate a function set when the default function set is fine.

Far worse, the aime libraries need to be initialized, the data types and the functions they provide need to be consolidated and when done, the libraries need to be disposed of. And that’s before starting to mention that the library objects define dependencies on other (library and non library) objects, and that these dependencies need to satisfied.

For that matter, improving on the interface and providing shorter paths to getting things done is a noted goal.

It is still worth noting that there are reasons for exposing all the joins in an assembled aime executive and that many applications will require access to some of them.

A new programming interface providing for a simpler aime application integration is currently being refined. While it does not aim at replacing the existing interfaces, it does provide an express way to assemble the various parts of the aime executive.

See Executive Assembler.

The new interface is still experimental and hence subject to changes.

The introduction of new referable and referring data types is equally complex at the very least. The types allows references to objects to be stored with data. Typical examples include containers, such as sequences and associative arrays.

See Referable Objects.

The design of such data types needs to thoroughly observe the mechanics of data references management, and by their very nature the requirements are quite imposing.