Rename source term variables consistently

[JoshuaLampert]

I renamed all source terms as s1, s2, ... for consistency and the d at the beginning of the variable names before indicated it's the time derivative q1 , h, ..., but it's only one part of it (for equations involving elliptic systems it's also the source term before applying the elliptic operator). So I think it was a bit confusing. See also #198.
On the way I also replaced some more expressions to avoid redundant computations and make the code a bit shorter.

[github-actions]

Benchmark Results

	main	c1b509f...	main / c1b509f...
bbm_1d/bbm_1d_basic.jl	13.8 ± 0.31 μs	13.7 ± 0.28 μs	1.01 ± 0.031
bbm_1d/bbm_1d_fourier.jl	0.217 ± 0.014 ms	0.248 ± 0.31 ms	0.877 ± 1.1
bbm_bbm_1d/bbm_bbm_1d_basic_reflecting.jl	0.0806 ± 0.00041 ms	0.0813 ± 0.00028 ms	0.992 ± 0.0061
bbm_bbm_1d/bbm_bbm_1d_dg.jl	0.0345 ± 0.00062 ms	0.0341 ± 0.00056 ms	1.01 ± 0.025
bbm_bbm_1d/bbm_bbm_1d_relaxation.jl	27.4 ± 0.49 μs	27.4 ± 0.46 μs	1 ± 0.025
bbm_bbm_1d/bbm_bbm_1d_upwind_relaxation.jl	0.0485 ± 0.00064 ms	0.0485 ± 0.00064 ms	1 ± 0.019
hyperbolic_serre_green_naghdi_1d/hyperbolic_serre_green_naghdi_dingemans.jl	4.15 ± 0.014 μs	4.09 ± 0.013 μs	1.01 ± 0.0048
serre_green_naghdi_1d/serre_green_naghdi_well_balanced.jl	0.202 ± 0.011 ms	0.204 ± 0.01 ms	0.989 ± 0.072
svaerd_kalisch_1d/svaerd_kalisch_1d_dingemans_relaxation.jl	0.147 ± 0.0061 ms	0.149 ± 0.0069 ms	0.987 ± 0.061
time_to_load	2.1 ± 0.016 s	2.09 ± 0.013 s	1 ± 0.0098

Benchmark Plots

A plot of the benchmark results have been uploaded as an artifact to the workflow run for this PR.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

📢 Thoughts on this report? Let us know!