dhil/phd-dissertation
1
0
Fork 0
mirror of https://github.com/dhil/phd-dissertation synced 2026-03-13 11:08:25 +00:00
Code Issues Releases Wiki Activity

Notes on why first-class control matters

Browse Source
This commit is contained in:
Daniel Hillerström
2021-05-18 23:45:20 +01:00
parent 8ecfadb94e
commit c28aef1c9e
1 changed files with 26 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

30
thesis.tex
View File

@@ -394,11 +394,33 @@ In this dissertation I am concerned only with
parameterised handlers which are a slight variation of deep handlers. parameterised handlers which are a slight variation of deep handlers.
\section{Why first-class control matters} \section{Why first-class control matters}
From the perspective of programmers first-class control is a valuable
programming feature because it enables them to implement their own
control idioms as if they were native to the programming
language. More important with first-class control programmer-defined
control idioms are local phenomena which can be encapsulated in a
library such that the rest of the program does not need to be made
aware of their existence. Conversely, without first-class control some
control idioms can only be implemented using global program
restructuring techniques such as continuation passing style.
\subsection{Flavours of control} From the perspective of compiler engineers first-class control is
\paragraph{Undelimited control} valuable because it unifies several control-related constructs under
\paragraph{Delimited control} one single construct. First-class control can be beneficial for
\paragraph{Composable control} implementing programming languages which have no notion of first-class
control in source language. A runtime with support for first-class
control can considerably simplify and ease maintainability of an
implementation of a programming language with various distinct
second-class control idioms such as async/await, coroutines, etc,
because compiler engineers need only implement and maintain a single
control mechanism rather than having to implement and maintain
individual runtime support for each control idiom in the source
language.
% \subsection{Flavours of control}
% \paragraph{Undelimited control}
% \paragraph{Delimited control}
% \paragraph{Composable control}
\subsection{Why effect handlers matter} \subsection{Why effect handlers matter}