There is no universal definition.  It depends on the application.

I was going to say that I couldn't come up with 10-wordword definition, but that I could feel it when I saw it.  I do like your definition.  Do you ever think there would be  a time when variability isn't the primary driver of quality?

If the variability doesn't matter, how is that related to quality?  You need some threshold or something.  But variability will probably play a role in the definition.

...the definition I put forward is better interpreted as 'My ability to produce acceptable product is highly proficient' rather than 'My product is of acceptable quality'.

Quality = c / variability  "I can produce product: reliably, reproducible, consistent with target specifications with a high degree of accuracy and precision, and with high likelihood of acceptability'.

Interesting discussion indeed. Let me read more about your exchanges! Naturally the smaller thhe variability of the results are, the better the quality of the measurements. But other parameters are certainly essential. Remember: small variability is not sufficient to ensure accuracy. Accuracy is at the end more important than low variability to obtain acceptability, although it is hard to prove accuracy if you have too little repeatability. Think also that customers need also "timely" answers. Have a good day.  

I'd rather agree with Stein. The suggested definition of [measurement] quality seems not to take into account a possible systematic bias of a set of results with low variability between them. In other words, in the case of measurement quality, one have to take into account the distance of a result from the true value (=accuracy) as one of the target specifications (another one is reproducibility). 

As to the "timely" answers... A "timely" (presumably of a lower quality) result still has to be associated with a reasonable uncertainty. Whether it is acceptable for the customer, depends on the requirements (=target specifications). 

The definition I put forward (Quality is inversely proportional to variability) is the general modern definition from the science of statistical quality control (Montgomery 2005). It refers to the general capability of ones process for producing conformable product - i.e. high quality and low variability implies one can implement the same process in the same manner reliably. Its written this generally so that an quality control can encompass any domain: production, service, measurement, healthcare, management, etc.

Following that definition, Montgomery describes dimensions of quality, including: performance, reliability, durability, serviceability, aesthetics, features, perceived quality, and conformance to standards.

Going specific to the measurement domain, and specific speaking about measurements (and not measurement processes and procedures) I'm in agreement that trueness and precision (probably better expressed in total as mean-square error and RMSE) are paramount.