Re: footnotes

On Tue, 9 Jun 1998, Alan G. Isaac wrote:

> One might use footnotes for many things, and there are many possible
> implementations. The basic function of a footnote is to designate
> some text as subsidiary.
> 
> I don't understand why html does not offer explicit footnote and
> endnote tags, both for the structure this offers to html documents
> and to facilitate translation between document formats. (The latter
> reason alone wd be reason enough, but there are clear structural
> reasons---reflected in the existence of footnotes---as well. There
> were discussions recognizing this and proposing to include footnote
> tags in html 3.0.)
> 
> Of course one can use make shift ploys: e.g., superscript a
> hyperlinked number. But i. this does not acknowledge the special
> structure of footnotes, and ii. this does not permit algorithmic
> translation *from* html to other formats.

You are right that there is no easy way to represent footnotes in HTML.

The specific semantics cannot be encoded in stone, but you can use a
pair of reciprocal links with REL/REV="Footnote" (or "Endnote") to show
the relation. The most obvious place for the elements in the document
strucuture would be the end of the document.

Rendering endnotes is easy - just include them at the end and link with
REL="Endnote" (with a link of REV="Endnote" at the endnote so you can
get back to the reference quickly).

In the case of footnotes, there are some possible cheats. In visual
media there is no difficulty since footnotes and endnotes will *always*
be at the end of the document - no pages, hence no page footers, hence
no footnotes. Printing causes some problems. Fixed positioning in CSS2
offers a way of creating page headers and footers in printed media, but
there's no way to generate them dynamically for each page printed. If
we go completely off the wall and talk of our dream user agent, then
client-side methods could be enabled in printed media and you could
have the dynamic generation there, but that would, in the end, reduce
the semantic meaning of such footnotes and hinder transcribability, as
you mention.

I don't know much SGML, but I suspect DSSSL would be able to handle
this. There is no element that conveys the semantics correctly, though,
and yes, the HTML 3.0 proposed syntax was rather nice. You can use
endnotes for now as described above, I suppose.

-- Stephanos Piperoglou -- stephanos@internet.com ---------------------
Visit HTML with Style at              http://www.webreference.com/html/
   Every second Thursday a new tutorial on HTML and CSS, plus much,
    much more... for those who like to author Web pages with Style
-----------------------------------------------------------------------

Received on Tuesday, 9 June 1998 04:17:30 UTC