ietf-822
[Top] [All Lists]

Re: MIME to Draft Standard

1993-01-21 07:13:32
From my point-of-view, I have a similar concern -- I want to be able to
extend the set of richtext commands that do not nest (and thus do not
require a matching "end command"), and to be able to tell by looking that a
particular command is non-nesting.  (Such commands, in my view, might not be
...

Keith,
  Thinking about this from an SGML perspective, there is an important
distinction.  Rather than try to explain the details right now and bore
people to death, what do you mean  by "non-nesting"?  Is the issue that
there are some things that inherently don't need end-tags, so you want
to leave those out?  Or do you have examples of things which need to be
expressed in a way that is inconsistent with a hierarchical or nested
model of the universe (e.g., that would have to span things that were
otherwise neatly nested in a variation of a GOTO)?

   If one defines a DTD that way, there are lots of circumstances in
which end-tags can simply be omitted even though they are implicitly
there.  In programming language terms, several (but not all) of the
mechanisms for this involve what we would call multiple closure-- the
appearance of a certain end-tag (or sometimes any tag at all) implies
closure of everything opened after the corresponding start tag (or by
the last start tag).  

Some of us hate the added flexibility this provides (too confusing to
users) and have built definitions that specify exactly what end-tags can
be omitted, require that they be omitted, and then use syntax in the
generic identifiers that make it visually obvious which is which.  Is
that something like what you are looking for?

   john

<Prev in Thread] Current Thread [Next in Thread>