W3C home > Mailing lists > Public > www-html@w3.org > May 2005

Re: separator abuse

From: Orion Adrian <orion.adrian@gmail.com>
Date: Tue, 31 May 2005 14:31:01 -0400
Message-ID: <abd6c801050531113158aa44dc@mail.gmail.com>
To: www-html@w3.org

Personally I would like to see <sep /> for lists as well at the same
level as <li>

	<li>Item 1</li>
	<li>Item 2</li>
	<sep />
	<li>Item 3</li>

But that's not the issue that really bugs me.

Tree hierchies make for lousy document modeling of everything but the
overall document structure. They don't work well for character styles
(they often overlap) or tables styles (this should need no

The everything has a container argument and there are no empty
elements is becoming a religion. From my perspective, there are good
use cases for empty elements. SGML got it right apparently. There are
uses for empty elements that are replaced by content that can't be
represented by plain text.

Also character styles (styles applied to a span of text) are applied
to a range of text and not in a hierchical way.

Table styles are just a mess when it comes to hierchical
representation, especially how it's done in HTML.

The religion that is forming around "no empty elements" and
"everything is represented by a child/parent/sibling relationship" is
taking HTML into a place where it's becoming even worse than it was.
HTML 3.2 was bad. I'm not denying that, but the response to it seems
to be taking the spec to a bad place, but on the other extreme.

<separator /> defines that the link between two blocks of text is
important. Even if we went down the annoyng route that is grouping
everything by the cause of the separator, it's almost impossible to
implement, especially if the author is dead. Decisions for things in
writing, be it specs or fiction, non-fiction or bibles is arbitrary.
Authors often don't know why they went a particular route. They just

While it would be super nice if documents would simpy mark themselves
up decoding our thoughts without the need for translation, the amount
of effort that would take would be monumental. The number of people
that a) actually care about this and b) can do it is very small.
Marking up a document is an onerous process. There should be a limit
to the requirements we place on the author. If we don't, the
technology will never be used. This is a usability point that has been
presented over and over again.

There is a limit to the number of elements that a user can retain in
memory; a limit to the number of rules; a limit to the number of
design design principles. This isn't a religion. There are no
steadfast rules or doctrines; only guidelines and statistics. Do
usability tests and I can tell you a lot of people will find that
<part> or <group>ing everything simply to get the usage of <separator
/> will tick them off.

Still no one has told me why I should care. The parts above and the
parts below the <separator /> don't need to be addressed
independantly, given <section> or <ul>. Let's let the user use what is
easy for them to use. It doesn't break presentation/semantics, it's
easy to use, easy to understand, and easy to read.

Orion Adrian
Received on Tuesday, 31 May 2005 18:31:07 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 February 2014 07:19:04 UTC