The syntax is as in Murray's document. Every NAME should correspond to a unique label within the document.
There is no scheme in HTML for specifying the type of a pointer (reference) to a data item, so CML uses the string to encode the type. The format is:
The postprocessing software can use the type information to
determine what to do. For example:
HREF="MOL:aspirin"
could be rendered either as an internal hyperlink to the molecular information
or as an inlined image. The choice could be left as options on the
postprocessing software.
The following addressing scheme is suggested:
From HTML 2.0 DTD:
<!ENTITY % A.content "(%text)*" -- <H1><a name="xxx">Heading</a></H1> is preferred to <a name="xxx"><H1>Heading</H1></a> -->