Re: <PAGE> proposal

Wilfredo Sanchez Jr. (tritan@mit.edu)
Sat, 23 Dec 1995 16:44:38 -0500


Date: Sat, 23 Dec 1995 16:44:38 -0500
From: "Wilfredo Sanchez Jr." <tritan@mit.edu>
Message-Id: <199512232144.QAA12546@calloway.mit.edu>
To: esr@locke.ccil.org
Cc: www-html@w3.org
In-Reply-To: <199512231600.LAA20805@locke.ccil.org> (esr@locke.ccil.org)
Subject: Re: <PAGE> proposal


It seems to me that (at the moment) a "page" correspond to an HTML
"document", so there is only one page per document.  What you want is
a single document with multiple pages, which is the same as mutiple
documents.  I can understand why you feel you need to decouple this
relationship, but perhaps it's not necessary (see below).  It's a
coincidence that when you send a URL to a server, it fetches a "file",
which is really just a matter of implementation.  I could write a
server than returned a document from memory, or generated on-the-fly,
as CGI scripts might do, and so on.  I hardly see a need to decouple
"file" from "page", as they are really not explicitly connected by the
HTML specification.

(I may be misinterpretting your message, but...)

So what we have is a tie between page and document, and this bothers
you.  Does <link rel="next" ...> and <link rel="previous" ...> really
not need your needs?  It's a simple matter to write a browser which
displays a document (== a page) and then at the end, you could press
the spacebar and magically, it continues onto the next document.
Browsers really should have next and previous buttons which correspond
to the <link> tags, apart from the "last and next page visited"
buttons.  But there is nothing to keep you from printing a "muli-page
document" which extends accross several HTML documents, or from
displaying linked documents together, via the behavior you described
or another.

	-Fred
	 tritan@mit.edu