Re: Portrait vs. Landscape (was Re: THEAD & TFOOT for columns)

E. Stephen Mack (estephen@emf.net)
Fri, 15 Aug 1997 15:25:31 -0700


Message-Id: <3.0.3.32.19970815152531.00772a90@emf.net>
Date: Fri, 15 Aug 1997 15:25:31 -0700
To: www-html@w3.org
From: "E. Stephen Mack" <estephen@emf.net>
In-Reply-To: <199708152113.WAA21531@imbolc.ucc.ie>
Subject: Re: Portrait vs. Landscape (was Re: THEAD & TFOOT for columns)

[To all subscribers of www-style and www-html: Per the
instructions sent to all new subscribers, please do not 
cross-post to www-style and www-html.  Many of us are
subscribed to both lists, and don't need to receive two
copies of each message.]

There is some confusion regarding what is meant in the current debate
using the terms "portrait" vs. "landscape" in page display.

Peter Flynn wrote:
>That does not make any sense.  Right is right and left is left,
>regardless of whether the horizontal edge of your window is longer
>than the vertical or vice versa. Most screens are landscape: they are
>wider than they are high. I'm really unclear what the problem is here.

Yes, there's some miscommunication going on.  The point made
by Neil Laurent on 14-Aug-97 as a response to Jeni Tennison's
proposal for COLHEAD and COLFOOT was:

> More importantly however, [the lack of column head and foot groups]
> points out one fundamental flaw with HTML.  Nothing in the language
> was meant for horizontal scrolling, you can look all over the place
> in the language to find errors if horizontal scrolling is required.

His point is regarding horizontal SCROLLING.  

Some Web authors use wide tables to create long pages that scroll
horizontally instead of vertically.  For example, see

"That Rope" by Jef and Gael Morlan
   http://www.microaero.com/j'n'g/3-25.htm
or "Why Don't Video Game Makers Make Games for Girls?" by Hi-D.
   http://www.hi-d.com/issue0/what_about_us/article_1.html

(These pages requires a visual user agent that can display wide
tables, such as Navigator 2 or later or IE 2 or later.)

The issue became confused when Neil Laurent extended the idea of
horizontal vs. vertical scrolling, by metaphor, to talk about pages
that scroll vertically as "Portrait" layout pages (which is the
default behavior for all visual browsers), and pages that scroll
horizontally as "Landscape" pages:

> HTML as far as I can tell was designed for Portrait layout and 
> pretty much lacks any capacity to do landscape layout.  Note the lack 
> of columns of text, no floating objects top/bottom, and some more.

This is absolutely correct.  All HTML specifications including the
HTML 4.0 draft make the implicit assumption that on a visual display,
the page will be oriented top-to-bottom (with one paragraph below
another, vertically), and that all scrolling on a page that is too
long to fit on one screen will be done vertically, not horizontally.

For example, there are numerous references to "horizontal" tabs,
space "before" (assumed to be above) and space "after" (assumed to
be below) an element.  In addition, the point was made by Walter
Kaye that there is a horizontal rule and not a vertical rule.

While HTML itself does not care whether a page scrolls horizontally
or vertically, the draft does seem to make the assumption that
pags will be scrolled vertically.  For example, there are statements
such as:

| Visual rendering of paragraphs 
| How paragraphs are rendered visually depends on the user agent.
| Paragraphs are usually rendered flush left with a ragged right
| margin. Other defaults are appropriate for right-to-left scripts. 
| HTML user agents have traditionally rendered paragraphs with white
| space before and after, e.g., 
[example deleted]

in the section on text (text.html), and similar statements in the
section that defines the table row groups (referring to a theoretical
visual agent rendering a table using vertical scroll bars).

Now, it's arguable to debate whether or not HTML should provide
equal opportunity for long pages that wish to be display using
horizontal scrolling instead of vertical scrolling.  It is not
merely a matter of "layout", as Liam Quinn argued.  Instead, HTML is
currently "hard-wired" to be vertically oriented--through the
deprecated alignment values, such as LEFT and RIGHT.

Since these attributes are all deprecated anyway, I don't believe
that HTML has to make accommodations for horizontal scrolling.
But the HTML draft should clarify that for visual user
agents, a vertical orientation is an implicit assumption.

Future versions of CSS *should* allow for horizontal scrolling,
however.  Perhaps this can be achieved by allowing authors to define
a canvas size which will force visual user agents to layout the page
using one orientation or another.

Currently, a Web author who wishes to create a horizontally-oriented
page must abuse tables.  There is no recourse in CSS-1 for
horizontal scrolling.  Further debate on how future levels of CSS-1
should handle this issue (which is unrelated to how a page is
printed) should take place on www-style.
-- 
E. Stephen Mack <estephen@emf.net>    http://www.emf.net/~estephen/