an overview of the current research in psychology of programming: the most interesting aspects is the concepts used as theoretical frameworks and the fact that the mental model hypothesis is gaining traction.
useful concepts:
The starting points is a borrowing of cognitive (task-oriented) concepts
Mental activities construct representations and operate on them
within a given context, elements are being perceived, processed through existing knowledge schemas in order to extract meaning.
representations are circumstancial constructions, in a context, and for specific ends. they are determined by tasks, functions, goals, etc. p.7 (knowledge builds on this but it is more of a long-term process)
question: if it is task-oriented, what’s the relationship with aesthetics (passive contemplation)?
working with the computer is comprised of the prescriptive (what the computer should) to the effective (what the computer actually does)
categories of software psychology:
this particular field is concerned with the relationship between (1) human representation and (2) the formal systems supporting this representation
data + algorithms (wirth)
the interesting concept is drawn from chomsky: surface-structure (textual) vs. deep-structure (conceptual)
there is a strange thing about linearity in programming: it is strictly linear but nonetheless jumps all over the place making it seemingly/effectively non-linear.
-> there is an entanglement of the plan, of the arc, of the tension (not so often in narrative text)
-> program texts are dynamic, procedural texts (exhibiting complex causal relations between states and events)
-> the understanding of program text is a general one, not applied specifically to a particular situation
they are unambiguous, while NL are anaphoric (variable names are, here, an overlap). software documentation and pseudo-code stands as a hybrid.
the first obstacle is that problems in programming are “ill-defined”, and that the types of problems affect the heuristics deployed by the individuals (e.g. data-heavy, procedure-heavy, varying requirements, etc.) and so there are 3 hypothetical approaches as to how to solve these problems:
the activity of designing programs consists in activating schemas
there’s a mutually-influencing triad of understanding the problem, researching it and implementing it.
the work of programming is the reconciliation of 2 domains:
theory of schemas
there are multiple kinds of knowledge: synctatic, semantic and schematic. it’s also known as a frame, taken from the AI field, so it’s broader than a pattern. a schema is then a mental model, different from a ready-made solution plan, which would tend to be too precise.
the difference between experts and novices is that experts have more meta, broader mental models. interestingly, the super expert is one who deploys more breadth than depth (what that means is that being able to bridge domains is more important!)
cf. design patterns (i have that book)
Aesthetics in code is visualizing patterns through text
starting with the fact that “understanding is constructing [useful] representations”, do we understand either as text or as problem?
in any case, the sources of information are: the text, the self and the environment (you could say is actually situated, within the text). once we have these sources, there is a process of inference between pieces that lead to understanding.
WHICH OF THE RULES OF PROGRAMMING DISCOURSE IS BEAUTIFUL?
WHICH ONE ISN’T?
the most probable model is the last one, the mental model, with a highlight on beacons, as thematic organizers, and a difference of experts in adding situational models.
-> program comprehension as problem-solving
p.101
the problem-solving approach focuses on the goal-oriented reading (i.e. given a text, we don’t understand the same thing/don’t have the same mental model if we have different goals) -> this seems to be the most dominating now mit study on reading code, neither as math or as language
main question: how is surface structure transformed into deep structure?
finally, difficulties for understanding: