Before we can talk about Computer Science we have to understand what this "science" thing is. And here it is:
Science is about modeling repeatable verifiable things which allows humans to recognize, predict and, if possible, control those things or their consequences.
Every word in this definition is important. Science only deals with something repeatable and verifiable. If something cannot be repeated, we cannot verify the model. If we cannot verify the model, it is not useful, because we don't know if it's usable, if it will work.
Notice, "usable", "work", not "truth". Science does not deal with truth. Science does not know if something si true ot not. Science deals with models that either work or not. If the model works, it is useful, it is scientific. If it does not, it's false, that is it does not work. Notice, that although science does not know the "truth", it's about the models that work, the science definitely knows "false". Here "false" means that the model does not work.
All mimsy were the borogoves,
And the mome raths outgrabe.
— Lewis Carroll. Through the looking glass
Are these lines true or false? None of that. We just don't care. Science has to be useful and it is only useful if it is about something that we care about. At least potentially.
But what does it mean that "a model works"? A model about something works if it allows us, humans, to recognize that something, predict that something, and, if lucky, control that something.
"Works" means something useful. It means that we can use that model. For something useful for us. And to be such it must be about repeatable verifiable things and allow us to recognize, predict and, if lucky, control those things. That's what science is for. To recognize, predict and control things that happen (repeatable!) in our lives.
By the way, the hypothesis of God is not scientific. Not because it's false. Not because it's true. Simply because it's not a scientific hypothesis. It's not something that science deals with. Nothing about God is repeatable, or verifiable, or predictable, not even to mention controllable.
Again, a model does not have to be "true", it only must work. It should be usable for what it is used for. Say, you are going to a party and to find your friend's house you are using a flat map. It does not matter that the Earth is round, all that matters is that this "not true" model of Earth gets you where you want, to your friend's house.
It also means that we may have several different models for the same thing. For example, we can measure a temperature using different scales - Kelvin, Celsius, Fahrenheit. They measure the same thing but give us different numbers. And all those numbers are "true" since we can use them. Just like with maps and globe, or different projections of maps (Mercator vs. Spheric) we just use whatever is more convenient at the moment. The temperature of a liquid gas is better measured in Kelvin while the fever in a patient is better detected using Celsius or Fahrenheit.
Again, scientific models are not religions, they can coexist at the same time. And if the scientific physics community supports the String Theory and oppresses all alternatives in terms of grants and publications, it's not because they are scientific. In fact, they act in that case religiously.
That's what "the science" is.
To model repeatable verifiable things which allows humans to recognize, predict and, if possible, control those things or their consequences.
That's all.
