search

isaqb-org / glossary

An extensive glossary of software architecture (and development) terminology. Explains the terms used and referenced in the iSAQB foundation and advanced level curricula.
https://leanpub.com/isaqbglossary
35 stars 16 forks source link

"Degree" / "Maß" #161

Open mikesperber opened 1 year ago

mikesperber commented 1 year ago

The ISO 25010 quality attribute entries mostly start with "degree to which" or "Maß, in dem", sometimes followed by "Grad". The term "degree" is meaningless, "Maß" is just plain wrong. Cf.:

https://de.wikipedia.org/wiki/Ma%C3%9F_(Mathematik)

This sloppy use of language goes back to ISO 25010. (Its predecessor, ISO 9162, AFAIK, had an extensive section on metrology, which is missing in 25010.) But it's really causing headaches for trainers and the foundation group, where arguments of the sort "are quality attributes always quantifiable" cannot be resolved.

mikesperber commented 1 year ago

I don't know a good solution to this: Ultimately, the 25010 terms are just umbrella terms for measures or properties (or "attributes" whatever that is) of a software architecture, where the measures or properties themselves are left unspecified. So maybe we could just prefix all those entries by "Umbrella term for properties concerning ..."

alxlo commented 1 year ago

"are quality attributes always quantifiable"

In this discussion I would like to refer to https://en.wikipedia.org/wiki/Level_of_measurement

So IMHO measurable things do not need to be quantifiable. And therefore I don't see where "degree" is misleading as it could refer to a category of values as well as an ordinal or cardinal value.

But I'm not sure I've got the problem that is addressed here. I understand that @mikesperber is not happy with the definitions of (many) quality attributes. But what should change? There is currently no term for "degree". So should we create one? Or should we have a go at all quality attributes where that wording is used?

mikesperber commented 1 year ago

So IMHO measurable things do not need to be quantifiable. And therefore I don't see where "degree" is misleading as it could refer to a category of values as well as an ordinal or cardinal value.

Sure, but for many quality attributes, it's not even clear how they should be measured.

But I'm not sure I've got the problem that is addressed here. I understand that @mikesperber is not happy with the definitions of (many) quality attributes. But what should change? There is currently no term for "degree". So should we create one? Or should we have a go at all quality attributes where that wording is used?

I personally think the latter. But maybe you can suggest a suitable definition of "degree".