How to deal with the frequent revisions of natural constants.

[timbrisc]

Issue migrated from trac ticket # 178

component: help | priority: minor

2015-11-13 19:49:45: [email protected] created the issue

---

Since UCUM was first published in 1997/1998, natural constants such as the Planck Constant were modified, and modified rather frequently.

||1986 ||6.626 075 5(40) ||×||10^−34^ ||J.s|| ||1998 ||6.626 068 76(52) ||×||10^−34^||J.s|| ||2002 ||6.626 0693(11) ||×||10^−34^ ||J.s|| ||2006 ||6.626 068 96(33) ||×||10^−34^||J.s|| ||2010 ||6.626 069 57(29) ||×||10^−34^||J.s|| ||today ||6.626 070 040(81) ||×||10^−34^||J.s||

And similar alterations happen to other constants.

The question becomes what UCUM should do about these alterations?

A. Should we track them and maintain UCUM up to date? B. Should we change the values silently? C. Should we maintain the history of the prior values?

• e.g., the prior values could be maintained by adding a year index to the unit symbol

The pros and cons of these decisions need to be considered.

• if we update, then past conversions may no longer be the same, i.e., they may seem to be errors in numbers saved over time.
• if we don't update, then those same "errors" persist undetected.
• if we come up with something very complicated, such as indexing everything by version or year, etc. it just becomes hard for the users to deal with.

My proposal would be to just update these constants from time to time and not provide any special way of tracking them. If it turns out that only few of these constants change, we might preserve older versions with year subscript index.