Expanding the "metric" part

[romain-jacob]

You might have discussed this before, but I am not sure...

I would like to argue for expanding the "metrics" into two parts:

• the physical dimension or performance dimension: that is what we measure. For example: carbon intensity of the electricity, gCO2e, bandwidth, time, water usage, ...
• the metric: that is how we look at that dimension. For example, for the carbon intensity, a yearly average is a different metric than an hourly average. For bandwidth, the median and 95th percentile are different metrics too.

I think it is important to add this precision because even if two studies consider the same physical dimension they may reach different conclusions if they use different metrics.

Thoughts?