What to they measure it against ?

They could have asked me and I would have done it for less cost in less time, and I'm always happy with the results!

The birds are singing, commuters on train were freindly this morning, the boss isn't worried about that deadline ( until noon Monday anyway! ) and El Reg has got it's hands on a suitable Viz-style story involving many, many references to the gentleman's front facing equipmnent! Roll on 5pm!

Perfect silicone balls, eh? Are they planning a male version of Jordan?

If you have a molar mass on your balls - you ought to see a doctor!

So "changes of as much as 50 parts per billion" are a problem, but "We are trying for an accuracy of two parts in 100 million."

That would be 2,000 parts per billion, or 40 times the granularity of the aforementioned problem.

What am I missing from the story which would resolve this increased margin of error?

"So 'changes of as much as 50 parts per billion" are a problem, but "We are trying for an accuracy of two parts in 100 million.' "

"That would be 2,000 parts per billion, or 40 times the granularity of the aforementioned problem.

"What am I missing from the story which would resolve this increased margin of error?"

Isn't that *20* parts per billion?

2/100,000,000 =

20/1,000,000,000

...or am *I* missing something?

"Scientists have long called for a more stable yardstick"

Firstly, the SI unit of length is the metre.

Secondly, what has distance got to do with mass ?

Someone has been seriously mixing their units !

;-)

My reading is that the Australian reference to accurancy is with respect to the perfection of the spherical shape, not the mass, whereas the French issues with accurancy are to do with changes in the material affecting the total mass.

Is this not apparent to anyone else?

Here I was thinking that a kilogram was equal to 1 litre of pure water @ 20° C which equals 10 cm³ of water @ 20°C. It's getting worse than imperial measures.

Isn't the whole idea here to replace the reference mass (held in Paris) with a new standard based on the other SI dimensions (in this case, length) together with a physical object like an atom of Silicon(28)? For this purpose the boffins propose to create an aggregation of Silicon(28) atoms that is indistinguishable in mass from the reference kilogram. It is helpful to have this mass of silicon be spherical, because it is necessary to determine its volume with great accuracy. Knowing both the volume of this 1kg sphere of pure crystalline Silicon(28), and the packing density of crystalline Silicon(28), it is possible to compute the number X of atoms in the mass of silicon. I'm thinking that this number X is something around 2.15E25 or thereabouts. An independent calculation can be made, based on Avogadro's number and the mass of the Silicon(28) atom, to confirm this number X. The new standard for the kilogram becomes, "a spherical mass of crystalline Silicon(28) consisting of exactly X atoms" and we have a new, reproducible standard of mass that is grounded in nature herself.

How does the roundness of these balls compare with the roundness of the gyro rotors that flew on Gravity Probe B?

Nothing wrong with mixing your units on a Friday afternoon.