Improved limits for certain parameters

[rozyczko]

The results of the discussion with Jos and Grace about the parameter limits. Using [0,inf] or [-inf,inf] is not a good idea since the fitting (if started too far away from the right solution) can lead to non-physical minima.
The agreed on initial values should be as follows:

• thickness/roughness from 50% to 200% of current/default value
• sld between -1.0 and 10.0
• scale between 0 and 1.0

These of course only apply to values not set explicitly by the user.

Addresses easyscience/EasyReflectometryApp#289

[github-actions]

This pull request does not contain a valid label. Please add one of the following labels: ['chore', 'fix', 'bugfix', 'bug', 'enhancement', 'feature', 'dependencies', 'documentation']