You are using the term "readable code" without clearly defining it.
I think "maintainable code" is far more important.
Some of the things that help make code "maintainable" are:
- Sound domain abstractions.
- Wise program decomposition.
- Highly cohesive, loosely coupled modules.
- Descriptive, explanatory, consistent and regular names.
- Minimizing the exposure of implementation details. Minimizing the use of global data.
- Minimizing the scope of variables, pragmas, etc. Avoiding action at a distance.
- Avoiding duplication (DRY).
- Encapsulation.
- Interfaces that are: consistent; easy to use correctly; hard to use incorrectly; easy to read, maintain and extend; clearly documented; appropriate to your audience.
- Useful commenting and documentation.
- Components that are testable in isolation. Comprehensive test suites.
- Avoiding unnecessary cleverness (encapsulating and commenting it when cleverness is required).
- Establishing a rational error handling policy and following it strictly.
- Logging effectively. Logging enough information to trouble-shoot customer problems without the need to attach a debugger.
- Checking the result of all file operations, API calls and external commands, and handling unexpected results.
- Consistent code layout.
- Avoiding "big arse" functions.
- Avoiding magic numbers.
Perl doesn't stop me doing any of this. Indeed, it nicely supports all of the above points IMHO. Note that a more subjectively "readable" programming language
does not somehow magically prevent you from violating any of the above points.
Personally, I have little interest in cosmetic issues, such as whether one prefers forced indentation (a la Python)
or considers Perl's regex syntax or use of
$ in variable names "ugly" or "line noise", for
that is just personal (subjective) preference and has little to do with writing maintainable code.
See also:
