Warp schrieb:

>> With a liter of pure water at exactly 4 degrees celsius, one problem 
>> you'll have is to exactly hit the 4 degrees celsius. Another problem is 
>> to /get/ really pure water, and /keep/ it pure. Yet another problem is 
>> that you'll have to define the exact isotopic composition of the water.
> 
>   I don't see how that is different from the current method, ie. measuring
> the weight of that one object at 4 degrees celsius.

The difference is that, with careful handling, the number of atoms of 
that one unique object doesn't change (not significantly, that is), nor 
does its chemical or isotopic composition, and so these aspects are 
perfectly irrelevant; and even the temperature is perfectly irrelevant, 
as it has no influence on the number of atoms of an individual object 
(unless you cool it down so much that some components of the air 
condense, or heat it up so much that part of the object evaporates) - 
while it does influence the results when trying to reproduce the same 
mass with a litre of water at given conditions.

>   Except that with water you don't have to rely on one specific object which
> is unique and there exists only one in the world.

Which is actually the advantage in this case.

> 
>> Then there's the shape of the container. You need to make sure that it 
>> /precisely/ holds 1 litre when it is at 4 degrees celsius /and/ filled. 
> 
>   Not much different from defining length in relation to the speed of
> light. If you want to measure it, you need precise timing and precise
> length measurements.

Yes, but in that case you need to measure only the distance between two 
points. But trying to do that with a whole container is a good deal more 
complex, as you have to measure the relative position of quite a bunch 
of points.

That's also the reason why the attempt to define the kg as the mass of a 
certain number of Si-atoms - "counted" by measuring the crystalline 
structure and dimensions of a certain chunk of Si - uses a perfectly 
spherical shape for the chunk. Otherwise, computing the volume of the 
object would be infeasible at the desired precision.

Now while this is a quite well-understood process by now, creating a 
spherical /cavity/ in a solid object is yet a totally different thing.

>   At least with the water the definition would be fixed to one single
> weight. Then it's only a question of how accurately it can me measured,
> which is no different from all the other units.

The accuracy of measurement even with state-of-the-art methods is quite 
poor though, which is precisely the point.

Also note that scientists /are/ working on the problem of getting away 
from the Prototype definition; they don't want to lose precision though, 
so water is /not/ an option for them.

Post a reply to this message