[madengr]

There is an alternate definition using a sphere of silicon with N atoms:

https://www.nist.gov/physical-measurement-laboratory/silicon...

What's really need though is a universal, stable over eons, single standard for time, length, and mass. I believe time is N cycles of an excited sodium (light) emission. Length is N wavelengths of that same emission in a vacuum. Mass would be N atoms.

So why are they not using a single element to define everything? Is it a matter of finding the proper element that is easy to excite and stable enough (chemically and atomically) over the long term? Sodium is very reactive and easy to excite. Silicon is probably the opposite.

[gumby]

> What's really need though is a universal, stable over eons, single standard for time, length, and mass. I believe time is N cycles of an excited sodium (light) emission. Length is N wavelengths of that same emission in a vacuum. Mass would be N atoms.

Turns out they know that :-)

The problem isn't definition, its the accuracy of reproducibility. You need a mechanism that can be built from scratch and can be believed to produce identical results. You can pull that off counting wavelengths of an unknown number of a known pure atom. Counting actual numbers of atoms is quite difficult. Typically we do it statistically...via mass! We could get it precisely by using a small number, but then actually measuring the mass (rather than calculating it) would be difficult.

[dtech]

A second (time) is already defined on cesium transition, a meter (length) is defined as a fraction of the distance light travels in a second. Both things which we can accurately measure for some time now, and which universal and stable over eons.

What is the problem that would be solved by switching to a single element?

[madengr]

I don't recall. I just read some article recently about trying to define time and length with sodium; trying to define everything with a single element that is very common. Of course it could also have been a 50 year old Asimov book.

[CupOfJava]

They reference that silicon definition later on in the article.