The stop ruler

A ruler with an end stop - a small raised lip at zero, square to the scale - seats repeatably against the pubic bone and removes the error of a flat ruler sliding past zero as you press. A plain flat edge can creep a fraction under pressure, and every fraction moves where zero actually sits.

What a stop does

The stop sits flush at the zero end and projects at ninety degrees to the scale, so pressing the ruler firmly means pressing the stop against the pubic bone rather than pressing a flat edge that can creep as pressure increases. Once seated, the ruler cannot slide further in that direction without the stop registering it, which is exactly the repeatability a flat-edged ruler cannot guarantee under the same pressure. The seat matters because the bone is the reference: in a study of 201 men measured by seven andrology specialists, Habous et al. (2018) found pubic bone-to-tip length "more accurate and reliable" than measuring from the skin. Carpentry and engineering rules with a hook or lip at zero use the same principle for the same reason: a hard reference beats a flat surface you are trusting to hold still. A folding ruler complicates this further, since joint play in a folding ruler can undo a good end stop's repeatability at the hinge before the reading ever reaches the scale.

What it does not fix

A stop makes the seating repeatable; it does not by itself guarantee the scale's zero starts exactly at the stop. Some rulers, stopped or not, have a printed zero set a little in from the physical reference point, which is a separate fault with its own name and needs checking on its own terms even on a ruler with a good stop.

Improvising one

A stop is easy to add if your ruler does not have one: a small block, a folded card, or a second short ruler taped square to the zero end all work, as long as the added piece sits at true ninety degrees to the scale - an angled improvised stop introduces the exact sliding problem it is meant to remove, just at a different angle. Bevelled scale edges solve a related but different problem, cutting parallax rather than sliding, and the two features are worth having on the same ruler rather than treated as substitutes for each other. This has nothing to do with how a photograph-based estimate is produced, since an image model has no physical zero to seat against - it infers a figure from the picture rather than pressing anything to bone, and nothing to do with a rating score, which is a position in a distribution of images rather than a reading taken from a seated ruler. A person's stated opinion needs no stop either - a human judge is offering an impression, not a measurement with a seating problem to solve - and neither does Rate Cock's rating, which was never a length reading to begin with.