comprehending interactive programs

[Raoul Duke]

(these papers might be old news around here)

I've done both oo/observer and reactive programming. Personally I do
find reactive easier to deal with overall. But going further, I wish
that we had somehow more reified representations (visually,
graphically, gui style, interactively, able to see it from different
angles) of the data & control flow of programs.


"Based on this feedback and on our experience, we believe that the
main effect we observed comes from RP making it easier to reason about
data flow i.e., points (i) and (ii) in Section 2. This conjecture is
supported by previous findings on program comprehension showing that
developers understand programs by building a mental model based on
control flow [34] and data and control flow play a fundamental role in
the way programmers develop a mental representation of programs [6].
Data flow determines whether a variable depends on another, a
fundamental aspect of comprehending reactive applications. Detecting
dependencies among variables is also a reachability question because
it requires to check whether a change to a variable propagates to the
other. Previous research [23] noticed that reachability questions are
extremely common for developers, they are (i) hard to answer and
time-consuming and (ii) error prone. These findings also provide an
explanation of our results"



https://www.semanticscholar.org/paper/An-empirical-study-on-program-comprehension-with-Salvaneschi-Amann/1bfeed28baddac67648f83e0576664a5640e8aa7

(via https://2020.ecoop.org/details/ecoop-2020-papers/19/Tackling-the-Awkward-Squad-for-Reactive-Programming-The-Actor-Reactor-Model)