Fractional coordinate precision

Property name: Fractional coordinate precision
Description: The absolute precision of a set of fractional coordinates, in fractional units.
Type: float

This describes how precisely the coordinates were stated by the source they came from, which is generally coarser than the precision of the numbers as stored. A source that writes a coordinate as 0.3333 claims it to about 0.0001 of a cell edge, however many binary digits the value later occupies. It is the natural basis for choosing a matching tolerance, a symmetry-detection parameter, or a grid spacing, in place of an arbitrary constant.

{
    "$id": "https://schemas.httk.org/defs/v0.1/properties/core/fractional_coordinate_precision",
    "$schema": "https://schemas.optimade.org/meta/v1.2/optimade/property_definition.json",
    "title": "Fractional coordinate precision",
    "x-optimade-type": "float",
    "x-optimade-definition": {
        "label": "fractional_coordinate_precision_core",
        "kind": "property",
        "version": "0.1.0",
        "format": "1.3",
        "name": "fractional_coordinate_precision"
    },
    "type": [
        "number",
        "null"
    ],
    "description": "The absolute precision of a set of fractional coordinates, in fractional units.\n\nThis describes how precisely the coordinates were *stated* by the source they came from,\nwhich is generally coarser than the precision of the numbers as stored. A source that\nwrites a coordinate as `0.3333` claims it to about `0.0001` of a cell edge, however many\nbinary digits the value later occupies. It is the natural basis for choosing a matching\ntolerance, a symmetry-detection parameter, or a grid spacing, in place of an arbitrary\nconstant.\n\n**Requirements/Conventions**:\n\n- The value is an **absolute** bound on the deviation from the intended value, not a\n  relative precision and not a count of significant figures. A coordinate near `0.0001`\n  and one near `0.9999` written to the same number of decimals have the same precision.\n- The value is **dimensionless**, being a fraction of a cell edge. Multiplying by a cell\n  length gives the corresponding precision as a distance; see\n  [`length_precision`](https://schemas.httk.org/defs/v0.1/properties/core/length_precision).\n- Where a set of coordinates is described by a single value, it SHOULD be the **coarsest**\n  precision among them, since the set as a whole is only as precisely stated as its least\n  precisely stated member.\n- Where a source states both a number of digits and an explicit standard uncertainty, the\n  value SHOULD be the larger of the two, that being the weaker and therefore the safer\n  claim.\n- The value MUST be strictly positive. A value of zero would assert exactness, which is a\n  different claim; `null` MUST be used where the precision is unknown or where the source\n  states an exact value rather than a measured one.",
    "examples": [
        0.0001,
        0.001
    ],
    "x-optimade-unit": "inapplicable"
}

Fractional coordinate precision (property)