/ GOOD PRACTICE, OVERENGINEERING

An example of overengineering - keep it WET

This week’s post is pretty short. I’ve already written about overengineering, but this adds a personal touch.

I had to rewrite my Jet Train demo to use another data provider, switching from a Swiss one to a Bay Area one. One of the main components of the demo is a streaming pipeline. The pipeline:

  1. Reads data from a web endpoint
  2. Transforms data through several steps
  3. Writes the final data into an in-memory data grid

Most of the transform steps in #2 enrich the data. Each of them requires an implementation of a BiFunction<T,U,T>.

These implementations all follow the same pattern:

  • We evaluate the second parameter of the BiFunction.
  • If it is null, we return the first parameter;
  • if not, we use the second parameter to enrich the first parameter with and return the result.

It looks like this snippet:

fun enrich(json: JsonObject, data: String?): JsonObject =
  if (data == null) json
  else JsonObject(json).add("data", data)

In the parlance of Object-Oriented Programming, this looks like the poster child for the Template Method pattern. In Functional Programming, this is plain function composition. We can move the null-check inside a shared function outside of the bi-function.

fun unsafeEnrich(json: JsonObject, data: String?): JsonObject =
    JsonObject(json).add("data", data)                                          (1)

fun <T, U> nullSafe(f: BiFunction<T, U?, T>): BiFunction<T, U?, T> =            (2)
    BiFunction<T, U?, T> { t: T, u: U? ->
        if (u == null) t
        else f.apply(t, u)
    }

val unsafeEnrich = BiFunction<JsonObject, String?, JsonObject> { json, data ->  (3)
  unsafeEnrich(json, data)
}

val safeEnrich = nullSafe(unsafeEnrich)                                         (4)
1 Move the null-check out of the function
2 Factor the null-check into a BiFunction
3 Create a BiFunction variable from the function
4 Wrap the non null-safe BiFunction into the safe one

We can now test:

println(safeEnrich.apply(orig, null))
println(safeEnrich.apply(orig, "x"))

It works:

{"foo":"bar"}
{"foo":"bar","data":"x"}

When I finished the code, I looked at the code and thought about the quote from Jurassic Park:

Your scientists were so preoccupied with whether or not they could, they didn’t stop to think if they should.

I’m no scientist, but I felt it applied to the work I just did. I realized that I refactored in order to comply to the DRY principle. When looking at the code, it didn’t look more readable and the code added to every function was minimal anyway. I threw away my refactoring work in favor of the WET principle.

There are two lessons here:

  1. Think before you code - this one I regularly forget.
  2. Don’t be afraid to throw away your code.
Nicolas Fränkel

Nicolas Fränkel

Developer Advocate with 15+ years experience consulting for many different customers, in a wide range of contexts (such as telecoms, banking, insurances, large retail and public sector). Usually working on Java/Java EE and Spring technologies, but with focused interests like Rich Internet Applications, Testing, CI/CD and DevOps. Currently working for Hazelcast. Also double as a trainer and triples as a book author.

Read More
An example of overengineering - keep it WET
Share this