Philosophy & History of MuSR code

Just a bit. This is a place for some speculations.

The killer of data analysis software is sustainability (i.e. who maintains it) In MuSR very good software packages, like <wimda>_, are mostly a one man effort. <musrfit>_ may be an exception, but it is certainly not an open sofwtare effort. Both have the great advantage of a very professional user as the maintainer (the guru), therefore what you need is there, although through the personal preference of the guru.

Many attempts to build a software as a professional collective fight principally with the reduced community of user, not justifying very large efforts, but also with professional programmers lacks of the feeling of real needs.

Some of the needs are a bit complex. The user should find it extremely easy to think of the model and to modify it to suit the physics under investigation.

A crucial part of this is global fits. They pose two problems: they are cumbersom to build and adjust; their goodness is not straightforward to assess. <musrfit>_ offers a very good solution, but admittedly it has a steep learning curve.

Another issue is ease of installation. Certainly CERN root is the source of minuit, hence a necessary reference for physics data fits. But it is a bit of a bloatware.

In this context herw we explore another solution seeking the following advantages

• Intuitive GUI approach, with attention to power-execution, presently from single fits to suites.
• Based on python, a present growing standard.
• Inserted in Jupyterlab, a growing standard for data analysis, and offering with that a Browser interface
• Easy to install, also on macOS and Windows.
• Easily documented by ReadTheDocs
• Implementing continuous integration [not yet ;)]