list-style-type default UL behavior (was Re: CSS3 : font-style)

>From: "L. David Baron" <>
>Subject: RE: CSS3 : font-style
>Date: Sat, May 29, 1999, 9:51 AM
> list-style-type: cycle(disk, square, circle); /* something like typical UL*/

I think at first glance the typical UL actually acts like:

UL { list-style-type: cycle(disk, circle, square); }
    /* swapped order of square and circle from David's example */

Until you get to the fourth (and deeper) level of embedded unordered list,
at which point, the list-style-type appears to stay "stuck" at third level's
square, which implies the following default behavior:

UL { list-style-type: disk }
UL UL { list-style-type: circle }
UL UL UL { list-style-type: square }

Of course, that doesn't mean it's *right*.

And why is it that one can specify an empty or filled circle (circle vs.
disk), but one cannot specify an empty or filled square?

So, two questions to the list:

1. What do people think of a new value on list-style-type (say, "block") to
indicate an empty square (The "square" value appears to indicate a filled
square currently - another alternative would be two new values on
list-style-type to explicitly indicate filled vs. empty squares, and to
leave 'square' as ambiguous/unspecified).

2. Since the default nested list-style-type is fairly useless at a depth of
four and greater, what do people think if the default nesting
list-style-type behavior were changed from the above grouping of three UL
rules, to the following:

UL { list-style-type: cycle(disk, circle, square, block) }

Note: This is not a proposal to add the "cycle()" notation to CSS - I'll
leave that to Mr. Baron - I am simply using his clever notation to indicate
the concept.

This new default rule has the following visual/aesthetic advantages:

1. It continues to distinguish nested list items at depths of four and
greater from their containing list items.

2. It visually alternates filled vs. empty glyphs for the list-style-type,
so that nesting depth is more quickly easily perceived. (filled circle,
empty circle, filled square, empty square, filled circle, empty circle,

Remember, this "new" behavior would be identical to the "current" behavior
for embedded unordered lists of depths one through three - the proposal only
changes what happens at depths four and greater (where the "current"
behavior is pretty lame anyway).

Perhaps I should phrase the questions a bit a differently - how offended
would folks be if a user-agent went ahead and implemented this behavior (the
new value of 'block', and the four value rotating list-style-type)?

Tantek Çelik
Internet Explorer, Macintosh Edition

Received on Sunday, 13 June 1999 13:42:57 UTC