This is the best article I could find about this: https://www.geosci-instrum-method-data-syst.net/8/227/2019/g...
This happens because it's almost impossible to write a paper that includes every single detail needed to reproduce your work and fit within any journal's page count limits. For metallurgy it's like trying to write down your family's double chocolate chip cookie recipe except your chocolate chips are atoms, your oven's temperature varies +-500C, and all of your ingredients can spoil when mixed in the wrong order, cooled at the wrong rate, heated at the wrong rate, aren't mixed with trace elements of other ingredients that came in your ingredients, etc.
I have no experience with this but I've heard this also applies to chemistry glass blowers: https://www.nmr.mgh.harvard.edu/news/20150127/nearly-lost-ar...