A third party passing judgement? (I am reminded of the "committee of disinterested citizens" in "Atlas Shrugged")
Who is qualified to pass judgement on Quality, Parsimony, and Beauty? My experience with the FDA says that third party input in ANY of these categories is of negative value, but that's just my "IMHO".
To pass such judgement would require knowledge of the problem space sufficient to pass such judgements. Seems as if the sponsor of the system (the "using" enterprise) is best equipped. A single doctor can tell me more than an army of federal bureaucrats about each of those issues.
Who is actually qualified to pass judgement on a model that is meaningful? Caveat Emptor.
I was not clear. IMO the 'system that does SE' produces a descriptive model of the problematic situation then produces a prescriptive model of the intended problem suppression system. Then the 'system that does engineering/construction' elaborates the prescriptive model then expresses it in technologies (make/buy components that are interoperable). Then deployment, adoption, assay and adaptation occur.
The users of the SE-produced models are the engineering of systems denizens. My question is, shall the burden of incoming inspection and acceptance of the SE-produced models fall solely on the system that does engineering of systems or shall an independent and objective third party pass judgement on the quality, parsimony and beauty of the respective models?
On May 3, 2012, at 9:27 AM, Mike Dee wrote:
The user of the model is the system that does SE. More formally, the user of the model is the enterprise that sponsors the project, or the delegates thereof.
It is the enterprise that must define its risk tolerance, and then must determine if the cost of information exceeds the utility of information (is cost of information is worth it?). Oops. There goes that "worth" word again.
Regarding the last sentence --- Is it really up to the User only or shall the 'The System That Does SE' include a Model Quality Assessment node? (I vote for the latter.)
Are there criteria for "model" such as a) reflects the truth, the whole truth and nothing but the truth, b) describes/prescribes the minimal requisite technologies and c) demonstrates emergent characteristics and properties. (I think we should establish criteria.)
Are the current tenets of IV&VT necessary, sufficient and lean? (I think not.)
On May 3, 2012, at 8:05 AM, Mike Dee wrote:
The value in a model lies in providing a way to conceptulize and bound the problem and its solution. It is up to the user of the model to determine the risk associated with its use.