A bit of history on the Outspec (output specification) list types. The FOSI
goes back to the early 90's as markup was really still evolving (who here
remembers SGML?). At one point the thought was both data AND style would be
included in the document instance. However as users began to foresee
different output styles this approach was quickly dropped.
However by this time various list styles had been identified and were
carried over to the OS. Typically these are not read and the numbering is
based on specific requirements for the output contained in requirements
documents. The embedding in the stylesheet approach makes for a more
consistent output.
The types of lists in the OS are a good example of different lists you can
put into your schema/DTD and thus style sheets. I would hesitate to make an
element for each list though, especially enumerated lists. Let the
stylesheet establish the list type based on where the list appears. Most
schema/DTD's I've worked with support three list types (one I really don't
consider a TRUE list). These are ordered or numbered lists, unordered or
random lists, and definition lists (I don't consider them a true list).
You could actually get by with one list element and set an attribute at the
list level to determine if the list is numbered or not. Then, again, let the
stylesheet do the numbering.
Another option that I have seen used is to use a 'label' attribute that
contains the string you want displayed (e.g., label="a", label="b" or
label="*", label="*"). For a random list, this may be okay, but a numbered
list requires either the author to enter the appropriate value or some kind
of post processing to populate the label attribute.
My approach is to make the authors job easier and the output consistent for
the users, so anything that will do that is what I tend to support and push.
Lynn
