Broad rules, ambiguous and vague, help to think about the subject, but do not serve as something definitive of what should be done. This rule does not say what should be explicit. The rule is broken all the time in many points of the language, in some no, of course. It is not always easy to find the point of balance of what should be explicit or not. If it is all or nothing gets complicated.
I remember once a teacher asked for a job to make a code where all functions were broken into the simplest possible operation, I questioned. I delivered on purpose a code that was practically an Assembly, each function only made a single operation as simple as possible, never two or more. So outside of doing something dirty, the right thing is to look at the context, gain experience, think about it and decide case by case.
If you take the rule to the last consequences you have to do what I just described with variables, each simple operation must be stored in a variable, will you do this? I think it’s silly.
One variable should be used primarily: a) to store a data temporarily to be used later; b) because it needs it longer or because it will be used more than once. Nothing prevents using the variable for other things, including a recommendation to use a variable to better document what you are doing and so avoid a comment saying what you are doing or leaving with nothing and it is not clear what this is. That’s what you’re talking about, but it’s secondary use of the variable.
In the example what did this variable add? For me nothing. It essentially reproduces the name of the method you are calling, does not document anything. What good does that do? Everything you do in the code by putting or taking out needs to have an advantage, even if it’s not functionality, you need to know how to justify why you did it, you can’t do it by doing it, or because someone said you should do it, that someone is not writing his code and it will not be that person who will suffer the pains.
You’re on the right track. Some people may disagree, but I lose respect for who does it :P Remembering that I’m talking about this case.
The second case presented in the question is already more complicated and would need to better understand the context. There already generates more controversy because the variable generates new information and can be useful to make more readable. I do not know if it is so necessary.
Python is a language of script then you have two things to think about, first that scripts shouldn’t be this complicated, so it shouldn’t be so necessary. On the other hand, it should be script performance doesn’t matter, so creating a random variable won’t hurt much. In many languages create a variable until some of the code, but not in Python and the creation of a variable has a higher cost than most languages (no script).
I look good on the wall there and I think saying go here or there falls a lot in opinion. I would use a different pattern depending on the language or type of the project. In Python I would not create the variable, but I would only use Python for very simple things and not for complex things as many people do, for me it would be the wrong tool.
If you’re a team follow what the team tells you, if you define what to do create a pattern and follow it (create less vague rules).
I use the first approach to reduce the line of code, because sometimes the line ends up leaving the area of the editor. In addition, the first approach is much more readable for both you and your team. However, there are cases where the second approach is better, such as when you have a small function with few parameters. It is much better not to explain the variable in this case.
– JeanExtreme002
In addition to the issue of the line being longer in the second case, you may eventually want to validate the variable
client_name
(to ensure for example that there is no invalid character). Then this validation would put you in a row betweenclient_name = ...
andwb = ...
. Of course, in this example it is no problem to modify the second version for the first one, but the idea is that having the code arranged for more lines simplifies when adding functionality at a specific point of the function.– jfaccioni
I think the odds don’t fit the question. The question is simply about readability, not about line size or any possible test. The approaches were examples to exemplify what I said in the first line: readability and code pattern. If the return of
_get_client_name()
is necessary somewhere else, to get in the way of 'rule' 79 characters per line or some subsequent test I think the variable should be made explicit. Anyway, thanks for the answers!!– Murilo Sitonio