Research Wiki | Misc / Languages

TODO: Better summary. And redo every section. EVERYTHING

RMPL is used by Titan, which follows the Model Based paradigm.

YAMPA originated from the Yale Haskell Group, and is a Functional Reactive Programming language. Also related to E-FRP (pdf) (same group and all).

Ada

I know nothing of Ada, other than it is old.

Apparently it has some useful features for robotics, most likely dealing with timing properties. At least it is probably easier to understand the the Haskell and Lisp derivatives.

A Language For Action (ALFA)

Designed by Erann Gat, and used in the ATLANTIS architecture, this language lets the user define a network of modules. These modules communicate using channels. Due to restrictions (for example, there are no functions (only macros), timing bounds may be placed on response times.

The language has syntax similar to that of Lisp.

TODO: paper link TODO: better description

Behavioral Programming

Behaviors are threads which can be composed into complete systems. Not sure where this paradigm falls: seems similar to subsumption at least for the terminology of "behavior" but also ModelBased due to the formalization (specifically the behaviors, and triggering events in a state-machine like notation).

YAMPA

Originated from the Yale Haskell Group, and is a Functional Reactive Programming language. Whatever that means. Also related to E-FRP (pdf) (same group and all).

Main source

Arrows

The main improvement seems to be "arrows" which are a generalization of monads. What are monads?

As far as I can gather, FRP has a few huge problems. Earlier FRP languages such as Fruit and Fran lacked arrows. Yampa doesn't, so it's better. Or something.

I think that this is going to take far more time to understand than it is actually worth. Especially since the code mentioned in the the tutorial paper is no longer available. If I find myself board with a day / weekend to kill, then maybe I will revisit this. In the meantime, I'm going to just have to assume that the project is dead for a good reason.

RMPL

Used by Titan, which follows the Model Based paradigm.

Might be interesting, but for now I am neglecting the Model Based work, as its problem domain is different. Basically, if you are operating in space, it is reasonable to model actions as atomic changes in state, which lends itself well to the reasoning that the model based approach uses. It is not, however, very helpful when actions are continuous (other work calls them activities to differentiate the two).