Loopers, Non-loopers and Control IPs

[emil14]

Context

As it turns out there are at least 2 types of components in FBP:

• Loopers
• Non-loopers

See the book.

You may have noticed that Figures 9.3 and 9.4 did not have the familiar "do while receive has not reached end of data" - this is because they are written in a style which assumes that they end execution after processing each IP. In FBP an inactive process will be invoked the next time an IP arrives at any of its input ports, so this kind of component will be invoked once for each incoming IP. A result of this is that it cannot maintain continuity across multiple IPs, but this is where the stack comes in. Since the stack is outside the process's local storage, continuity can be maintained across the invocations using the stack. This style of component is called a "non-looper", as opposed to components written in the "do while receive has not reached end of data" style, which are referred to as "loopers". This is not an externally detectable attribute of a component, but just depends on when and how often the component decides to end processing - as long as there is data to be processed, it will continue being reinvoked.

Problem

1. Figure out why are there these 2 types
2. Is it possible to solve the tasks without it
3. What are control IPs and why do we need them

---

Highly likely related to #35

[emil14]

It's not clear why this issue exists