Regexes, where is the problem?
Jan 20, 2019
There is a famous regex quote that I’ve ran into a lot lately:
Some people, when confronted with a problem, think
“I know, I’ll use regular expressions.” Now they have two problems.
This phrase was said by Jamie Zawinski, it became so popular that eventually programmers started to use it as a dogma, on Jeffry Friedl’s blog you can find some context on its origin.
Since this gave me some thinking I want to put my 2 cents on this argument.
Regex is a tool for finding patterns in words and eventually modifying this patters with other words. So if you’re either grepping through log files, checking some xml attribute to match a pattern, or anything like this it’s absolutely the most convenient tool.
But this subject inspired me a more general reasoning.
Programmers build software to resolve specific problems, softwares themselves are tools.
We rely on tools in development to make the process of building easier and more enjoyable.
The languages we use are interpreted in ways to give them some meaning, without parsing these would be useless, and these too are tools.
And all this is an example of job modularization.
When you solve problems one of the main activities is to split the issues into small blocks, and this is the way programmers usually approach their work;
dividing large topics into many smaller has lots of advantages, it becomes easier and maybe associable to something that you have already covered.
Since strings are one of the most unpredictable inputs you can get,
doing complex character manipulation in one step (as regex permits you to do) easily attracts critics.
Seen in this specific perspective regular expression is a low-level tool to approach a high level task,
you can solve the issue but mostly won’t get to reuse what you wrote,
you’re adding a dead-end to your codebase for an ‘ad-hoc’ case.
Generally, when designing an application, this is something you want to avoid.
Caring for the architecture to be maintainable, clean, and extensible,
keeps technical debt low, on the long run that’s something valuable.
But on the other side there are places where regular expression are invaluable,
and fit perfectly in the application structure so without context it’s hard to really take a part.
In conclusion this is something more general than just regexes,
there always is a better alternative solution,
it may sometimes not be worth it,
but it’s good to keep in mind that there is.