Is it argued that processing capabilities are superior to other seismic processing software or that it is easier to use? Does it get recommended for use because of cost or its wide use?
The packages are open source, so any user can look at the code, understand how it works (or why it doesn't) and even produce reproducible research so other geophysicists can test the results and use the methods. A great way to learn about seismic processing is indeed by reproduce earlier studies.
A lot of new processing methods are developed within the Seismic Unix and Madagascar groups, and they are great sources if one needs a special tool for a special problem.
The packages are free and are widely used by universities, but they can also be useful for processing companies. It is easy to share or sell projects if they are built in code that everyone have access too.
If the processing capabilities are superior or not is generally about opinion, but at least the results can compare well with commercial packages. Sometimes they can be used in the same workflow, e.g. the geometry might be easier to do with a GUI but processing might be just as easy in a script.
My own experience is that the learning curve is rather steep in the beginning, but once you understand the basic tools, Seismic Unix and Madagascar are just as easy, or even easier to work with. Especially if there is a problem or you need to do something unconventional with the data.