Wikipedia:Mirror test

This essay, Mirror test, refers to running tests, such as for template operations, based on incorrect, or invalid, test data, which can give the appearance that there are defects, or perhaps nothing might appear wrong, when in fact, the real errors are in the test data instead. The essay title "Mirror test" relates to the main mirror for the Hubble Space Telescope, which was carefully polished to incorrect mirror specifications and passed tests against the incorrect test data. When running tests, beware any tests run against live articles, which might have errors in the markup language, rather than assumed to be "perfect" articles to be used for comparison tests.

Title phrase
The essay title, "Mirror test" is related to the production of the main mirror for the Hubble Space Telescope, which was carefully polished based on incorrect mirror specifications, making the mirror "perfectly wrong" in operation after it was deployed in Earth orbit, and given time to settle. However, because the mirror carefully matched an incorrect focal length, the overall fix was to change the other telescope lenses to compensate for the mirror's incorrect, but precisely polished surface. In December 1993, astronauts on the Space Shuttle Endeavour went on the first Hubble repair mission. During the repairs, conducted as spacewalks from the Space Shuttle, the corrective lenses were installed, and other upgrades were made, to not only "correct" the telescope vision, but also enhance the capabilities beyond the original plans (see also NASA engineer: Frank Cepollina).

After lengthy investigation, optical experts concluded that the original, incorrect mirror specifications could have been detected, sooner, before the mirror was incorrectly polished to the wrong size, by running some typical optical calculations, as an easy mathematical sanity test that would have caught the invalid specifications at the outset of telescope production. Meanwhile, during the months that the flawed telescope transmitted "blurry" pictures to Earth, some astronomers were able to mathematically "correct" the fuzzy images to retro-engineer more realistic images, based on the unfocused data.