This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
SML in some places says that SML references must be derived from sml:refType, implying a requirement for PSVI. In other places, it states that SML references are recognized by the presence of sml:ref="true"/"1", which has no PSVI requirement. Further, SML-IF states an intention to change to allow recognition of SML references without PSVI, although it does not currently have this in place. Does SML really need to specify that all references must be derived from sml:refType, or could the looser form be used (any type that requires sml:ref="true"/"1" at the type level MUST be an SML reference according to the loose definition, regardless of its derivation hierarchy). Related: http://www.w3.org/2005/06/tracker/sml/actions/31
We decided at Toronto F2F that sml:refType is no longer needed. Any complex type can be used as a reference type and targetXXX and acyclic constraints can be defined for it. These constraints will only be validated for instances of the complex type for which sml:ref="true" or "1"
As agreed during the SML f2f meeting, the sml:refType have been completely removed from the SML schema specification. According with the new definition, SML references are elements that have either sml:ref="true" or sml:ref="1" specified