Simon Krughoff noted that the Firefly image stretch dialog enforces that the "min" and "max" values in the dialog must be within the range of actual values in the image.
This was undoubtedly aimed at providing support to users in not entering nonsensical values. However, there are valid use cases for entering ranges that fall outside the range in the particular image being displayed.
The most obvious case would be when setting up a stretch for a set of images from the same instrument, each with somewhat different maximum actual values, so that a common value in every image is displayed with the same color in every image. (Imagine, say, images with a potential maximum - saturation - value of 4000. For different spots on the sky, the actual range in a particular image might not get anywhere close to that - it might be 17 to 850 in one image, and 16 to 3500 in another - but I would like "521" to display as the same color in each one.)
It's clear that this should apply when the "units" menu is set to "Data". It's not clear that it's useful when using "%", but that could be discussed, and it might be quite confusing to users to have limits applied in one case and not the other.
This argument arose from LSST but is pretty clearly applicable to, e.g., looking at a set of WISE images in IRSA as well.