html-wg@oclc.org
Date: Sun, 11 Sep 1994 18:35:56 -0600
From: html-wg@oclc.org
Message-id: <199409120035.SAA27316@acl.lanl.gov>
Apparently-To: <html-web@acl.lanl.gov>
Apparently-To: <html-wg@acl.lanl.gov>
Apparently-To: <rdaniel@acl.lanl.gov>
Peter Flynn writes:
 > Terry writes:
 > 
 > > So Sidebar is perhaps a kind of block-linked-to-document rather 
> than block-linked-to-point, perhaps something else.
 > 
 > Ping! Sorry, but you pushed a button there. Conceptually, a sidebar is
 > basically a footnote that gets printed in the margin. I see no reason
 > to treat it any other way than as a special case of <footnote>, with
 > positional _recommendations_ in an attribute.
 > 

 I agree with Peter here.
 Marginal notes, sidepar, changepar, etc should be treated analogously,
 without undue emphasis on how a visual formatter displays them.
 The fact that a footnote is printed at the bottom of a page, and a marginal
 note printed on the margin is more a visual layout decision, rather than any
 significant difference in the logical structure represented by these
 constructs.

 For instance, when AsTeR renders footnotes, I had a rendering rule that
 resulted in the footnotes being spoken on the right speaker alone (AsTeR uses
 stereo) at a slightly faster speech-rate. In a way, you could then say that
 footnotes went into the margin.

 The advantages of encoding constructs like sidepar ina layout independent way
 are numerous; a browser can choose to render the object at a time most
 convenient to the user.

 For example, AsTeR provided a "float" rendering rule for footnotes, which
 allowed footnotes to "float" to the end of the rendering of the enclosing
 paragraph, section, or chapter, (with this choice entirely left to the
 listener).

 If a footnote or sidepar were to be too closely "bound" to the point of the
 text where they occur, such flexible rendering would be impossible.