(Posted on 2022-04-25)
Last week, I watched a video about coding and Docker configuration on a TV. I couldn't read any line of code. Then I thought about visually impaired people. How do they code? Every student in Thailand must learn to code. I presume the situation is similar in every country.
People widely use text-to-speech services these days. I cannot find any text-to-speech service for reading source code aloud. Let's assume we have a modified version of a text-to-speech service.
So I looked at source codes in different programming languages on the CodeRosetta website. I perceive Python code blocks by their visual structure purely. To read Python source code, I have to encode its visual structure, namely indents to sounds. Reading a nested code block won't be easy to understand. For example, reading twelve leading white spaces aloud will be very strange. In Lisp, reading open parenthesis and close parenthesis is more straightforward, but I will forget which parenthesis. So the best form of code blocks is in QuickBasic, which has different keywords between different kinds of blocks. For example, FOR with NEXT, and IF with END IF. Later I got a comment from Lemmy.ml, which told me that Ada also has different keywords between different kinds of blocks. Another idea from Lemmy.ml is the reader must convert the Python code block into a similar form as Ada or QuickBasic before reading.
MBasic refers to code by line numbers instead of code blocks. However, by listening to five lines, I forgot the line number. For example, when I heard gosub 70, I forgot what was at line 70.
In X86 Assembly, a programmer labels only the line that the program will jump to it. So X86 Assembly code looks much better than MBasic.
Still, coding in X86 Assembly can be exhaustive in many cases. For example, X86 Assembly doesn’t support recursion. Writing quick sort in X86 Assembly can be too difficult for learning to code.
Haskell doesn’t rely on code blocks. However, reading the symbols, for example, >>= is challenging. Prolog’s symbols are easy to read. For example, we can read :– as IF. Anyway, the Prolog programming paradigm is different from the mainstream one now. So Erlang, whose syntax is similar to Prolog, is a more practical alternative.
In brief, Erlang is a practical, less visual-centric program language. Because it mostly relies on names instead of code blocks, reading names aloud is much easier than reading code blocks aloud. Furthermore, the Erlang programming paradigm is more mainstream now.