Based on discussions in HowCanSomethingBeSuperGreatWithoutProducingExternalEvidence:

Some believe that small examples are insufficient to demonstrate the full power of certain concepts. They often suggest that it is the interaction of powerful concepts that creates the large benefits, not each concept in isolation. This raises some questions:

* Couldn't small examples show two or more interacting concepts also?

* Can one show a portion of a large example where multiple concepts interact and explain why they make things significantly better? 

* How large does the example have to be in order to show clear benefits?

* I tend to divide large applications into smaller task or event modules that communicate with each other mostly through a database and a handful of parameters. Couldn't an example be built around this concept such that only one task is presented in an example, or does task-wise-division ruin the ability of your GoldenHammer to work properly? If so, why doesn't your GoldenHammer work under the task model? (Some have suggested that the task model is fundamentally flawed, but have not made a good case in my opinion. See FundamentalFlawsInProceduralDesigns)

----

One doesn't have to demonstrate using just entire small apps. Examples/scenarios can also demonstrate '''portions''' of medium and large apps. The context can be described. With enough scenarios of a wide variety, complexities found in a large application can usually be approximated sufficiently to demonstrate the power or utility of a technique. The reader or skeptical party may also submit scenarios of their own if they feel that non-representative issues or contrived conditions are being used. This reduces the burden of an example from proving everything to everybody, but rather focuses on issues of concern of the other party, as encountered. There's no use arguing over something that is not an issue of contention. Ultimately, the reader will judge whether a scenario is realistic or if it addresses their own particular concerns.

----
See also: SystemSizeMetrics, ArgumentByLabToy

Indexing keywords: "Toy example", "programming contest"
----
CategoryExample, CategoryEvidence