VOTable-1.3 Erratum 2 - Permitting F0 in precision attribute in FIELDs
Author: Thomas McGlynn -Operations Interest Group Chair
Date: 22 October 2017
The following set of Errata amends the VOTable-1.3 (Ochsenbein et al. 2013) recommendation.
Rationale
The precision attribute is an optional attribute on floating fields which can be used to indicate the actual precision of the values. It is intended for clients displaying floating point values and as discussed in the standard in section 2.3 it is intended to be equivalent to the fractional field in the FITS TDISP keyword. In section 4.2 the definition of the precision attribute is given as:
The
precision attribute is meant to express the number of significant digits, either as a number of decimal places (e.g.
precision="F2" or equivalently
precision="2" to express 2 significant figures after the decimal point), or as a number of significant figures (e.g.
precision="E5" indicates a relative precision of
10–5).
There are many cases in which we may wish to store floating point values, but anticipate that these data will have integral precision and accordingly the FITS TDISP field (and the FORTRAN formats on which it in turn was based) allowed for values like F10.0. The historical precedent and text of the standard strongly suggests that the standard was intended to allow precision values of '0' or 'F0'. This erratum thus applies only to the schema associated with VOTable 1.3 not to the standard itself.
Change of standards text
No change is anticpated for the text of the proposal. The only change is in the schema associated with the precision attribute.
The proposed change in the schema is replacing the definition for the precType:
<xs:simpleType name="precType">
<xs:restriction base="xs:token">
<xs:pattern value="[EF]?[1-9][0-9]*"/>
</xs:restriction>
</xs:simpleType>
with
<xs:simpleType name="precType">
<xs:restriction base="xs:token">
<xs:pattern value="[EF]?[0-9][0-9]*"/>
</xs:restriction>
</xs:simpleType>
or some other equivalent change.
Impact Assesment
The only effect of this change is to made certain files that were currently considered invalid, to be valid VOTables. Conceivably some software has been written to only handle non-zero precisions.