AUTHENTICATE
What to Authenticate?
Authenticity is the quality or condition of being genuine or trustworthy. Decide what it is you want to authenticate – is it a heuristic, an algorithm, a procedure, a principle, a methodology, an explanation, or what? For good results it is advisable to clarify what you do and do NOT want to authenticate. Before proceeding, develop or acquire the wherewithal to determine the authenticity of whatever you are focusing on – and remember that history, context, prospects, and implications may all be involved.
Why to Authenticate?
The questions of whether or not something is “authentic” and why that matters, should be settled before devoting time to a task the outcome of which could be trivial anyway (unless trivial pursuit is the task at hand!). There are always many choices as to how to spend one’s time, and what to do next, so look to the choices that are worthwhile, and if authentication of something in particular is NOT worthwhile, choose something better.
When to Authenticate?
Since authentication requires enough evidence to make a reasonable inference, the time to commence an authentication depends upon the availability of that evidence, AND then sufficient time to process that data. This is another judgment call – the phenomena in question will not carry a sign indicating when sufficient duration has occurred to enable and support an authentication – you have to remember, guess, or get guidance.
Where to Authenticate?
Since authentication is a separate consideration from functionality, and the process of authenticating could interfere with functional procedures, authentication should also be done “outside of the thick of things” (or by a neutral [but fully informed] observer). This may not be easy to do, since some authentications are performed “in mid-stream” so as to authorize continuance or discontinuance – plan ahead better than this for next time.
How to Authenticate?
Look for indicators rather than measures, because measures are both more comprehensive and more data-intensive, but indicators summarize highlights, and that is sufficient for authentication. Statistical sampling techniques get better press than intuitive “eyeballing” but often a “quick and dirty” impression can be just as “correct” as more elaborate procedures: one of the Laws of Forecasting tell us that Methodology is Over-rated.
References
