Development of software is not science. Mathematical work on algorithm could be called science, in the sense 'mathematics', but very little part of programmers' work involves such mathematical thinking.

I'm sorry, but you are deluding yourself. Like many technical people you take comfort in objectivity, but science is expensive and require stable foundations. You need something as unyielding as the laws of physics to make real scientific progress, and even then things get really messy fast when you look at things like nutrition where variables are difficult to isolate and experiments hard to design.

Look at the trouble Economics has being taken seriously as a science. Now consider how fast hardware/software moves compared to economic change.

Software engineering methodology is not tractable to science, there is no signal, only noise. If you disagree, don't blame it on me, show me the results and how you apply them.