Jeremiah Flaga My thoughts and experiences on programming, life, atbp.

Programming by wishful thinking and why we need to read

Category:

Having been influenced by the idea of TDD (even though I have never done it before) and the idea of “respecting levels of abstraction”, (and the Clean Code book of course), in my recent tasks at work, because we do not have tests, I just make myself imagine that the UI layer (the Presenters, Controllers, etc.) contains the tests.

Then I try to start writing code on the user interface level of the app (the Presenters, Controllers, etc.).

I then ask myself, “At this layer in the app, how do I want my classes and methods/functions to look like?”

Then I try to create functions with as little set of parameters as possible – only those parameters that I think are needed at this particular point in the app.

For example, at the level of the user interface (the Presenter, Controller, etc.), I start to write something like this:

bookmark(item); or item.bookmark();

… to bookmark an item.

Or this

int steps = fitbitHelper.getSteps(date);

…to get number of steps from Fitbit for a specific date.

At the user interface level, I want to avoid doing something like this:


int steps = fitbitHelper.getSteps(date, acessToken, etc.);

...because at this point, I do not want to care about "how" I get the number of steps from Fitbit. At this point, I only care about getting the number of steps -- the "what".

When they look good already, I start to implement them.

public class FitbitStepsHelper {

    ...

    public FitbitStepsHelper(..., ...) {
        ...
    }

    public int getSteps(Date date) {
        // ...
        //
        // if access token is expired
        //     refresh access token
        //
        // get steps from fitbit using access token
    }
}

(Note: The real code gets the steps from Fitbit asynchronously using RxJava, but I did not put asynchronous code above so that it can be easily understood. )


Then… Last night, while reading Chapter 10 of “Growing Object-Oriented Software Guided by Tests”, I learned that someone (or sometwo) has already named this kind of process before.

Abelson & Sussman named it “programming by wishful thinking”.

This reminds me that there is nothing new under the sun. When [we do not read that much and] we think we discovered something new, we have to remember to assume that someone might already have discovered it before, and even formalized it or have already given it a name.

This also reminds me of one of the reasons why we need to read (or watch lectures):

The problems that we encounter today might have already been solved by someone else. We don’t have to do everything by trial and error. We have to learn from the experiences of others.

(If you have comments about my English, and other things, please tell me : smile:)

Comments