Separation of program state and data state is easier in custom C/R. This allows modification of program to find bugs and reuse the same data, which is currently more difficulty in automated C/R solutions that save program state. This is an active area of research and development for application-level C/R, but is not there yet.
Of course, all this assumes the bug hasn't been corrupting the data/analysis all along.
Separation of program state and data state is easier in custom C/R. This allows modification of program to find bugs and reuse the same data, which is currently more difficulty in automated C/R solutions that save program state. This is an active area of research and development for application-level C/R, but is not there yet.
Of course, all this assumes the bug hasn't been corrupting the data/analysis all along.