Re: HTML+ and browser functionality

Dave_Raggett (dsr@hplb.hpl.hp.com)
Tue, 29 Jun 93 17:13:55 BST


> In a previous message, you
> mentioned using the GROUP tag also for this purpose. Can you
> explain use of the GROUP tag? I'd like to use the GROUP tag to
> is define a set of links shared by a set of documents, such as
> chapters all containing links to a table of contents, or a glossary.

If you have a number of chapters (with one document per chapter) which all
share same table of contents, then one approach is to insert a link in each
chapter to the table of contents. This is ok, but a pain to maintain. It
would in many cases be much nicer to define the structure of the book in
one place, with pointers to the table of contents and the various chapters.

The GROUP tag can be used to define hierarchical groupings of tagged elements
such as chapters, appendices and so on. When you follow a link into a book
defined in this way, the browser can keep track of the context, and show
menu items or toolbar buttons for the table of contents etc. The idea needs
to be thought through in more detail, and is still a bit rough ...

e.g. in the document book.html

<GROUP type="table of contents">
<TITLE>Your Dog! </TITLE>

<A REL="partof" href="preface.html"> Preface </A>

<GROUP type="chapters">
<UL>
<LI> <A REL="partof" href="chap1.html"> Introduction </A>
<LI> <A REL="partof" href="chap2.html"> So you want a dog </A>
<LI> <A REL="partof" href="chap3.html"> Your new arrival </A>
<LI> <A REL="partof" href="chap4.html"> Training </A>
...
</UL>
</GROUP>

<GROUP type="appendices">
...
</GROUP>
</GROUP>

Rather than simple lists of elements, it seems nicer to allow a range of
markup so that the hierarchy can be viewed directly. The REL="partof"
attribute indicates that the associated document "belongs" to this group,
rather than simply a referenced work. You can therefore build up hierarchies
extending across a collection of documents.

The order in which such parts appear in the markup defines a browse sequence
and imply NEXT and PREV links when viewing these documents. This is
independent of the use of the UL element as seen above.

Note that a given document could be part of several different books. The
appropriate link to follow when the user presses NEXT depends on how the
current document was reached. Perhaps we need to extend the A tag to allow
authors to specify which context is intended for ordinary cross references.

Some more thought is needed on exactly how the browser knows which LINK
elements are implied, i.e. what menu items to show for a given document.

Regards,

Dave Raggett