3.6. Combining Marks vs. Non-spacing Marks

It was the original intention in the merger with 10646 to identify the terms "non-spacing marks" (from Unicode) and "combining marks" (from 10646). In the final editing of 10646 it became clear that there are some differences between these two terms that must be made clear.

All non-spacing marks are combining marks; however, there are some combining marks that are not non-spacing marks. The difference between the terms is that combining marks include Indic matras as well as non-spacing marks. The Unicode Consortium views this as an defect in the preparation of the document, and is filing a defect report. However, in the meantime, implementors of Indic characters should be aware of some subtle differences in terminology:

The derived terms such as composed character sequence (Unicode) are not identical with similar terms such as composite sequence (10646). Composite sequence also includes any base character followed by any sequence of non-spacing marks and Indic matras (possibly mixed).
In 10646, "if a combining character is to be regarded as a composite sequence in its own right, it shall be coded as a composite sequence by association with SPACE." This implies that an independent Indic matra (without preceeding SPACE) is not to be regarded as a composite sequence; that is, a stand-alone Indic matra is not a composite sequence. This is compatible with the Unicode interpretation, which does not require a SPACE before a matra.
In 10646, combined marks "are intended to be positioned relative to the preceeding base character in some manner." In practice, this condition cannot be met by implementations of Indic, since the ordering relationships among matras are more complex than for simple non-spacing marks. However, since the standard only indicates the intention, the Unicode interpretation is also compatible with the requirements of 10646.