Allowed nesting of elements in HTML 4 (and XHTML 1.0)
Legend
- An uppercase word stands for the corresponding element.
(Note that by
XHTML rules, element names must be written in
lower case, e.g.
<html>
, not
<HTML>
.)
- A lowercase word is a term which
describes a collection of HTML elements.
- Each entry is followed by a list of elements which
may appear within the elements specified by the entry.
If there is no such list, no nested elements are allowed.
This means that only text (#PCDATA, see next item) is allowed inside
the element; but if the note (empty) is given, it means that
no content whatsoever is allowed.
However, for flow, inline,
block,
OBJECT, and BODY
the allowed contents are described separately under the main entries for them.
- #PCDATA means "parsed character data", which is plain text
(without HTML tags,
but "escape sequences"
such as
ä
and ä
are allowed)
- CDATA means "character data", which is plain text where even
"escape sequences" aren't interpreted;
for a much better explanation, see the article
CDATA Confusion
by Joe English
- excluding ... means that the element must not contain
any of the listed elements, directly or indirectly.
Nesting rules for HTML 4.01 Transitional
HTML
-
HEAD
-
TITLE (required)
-
SCRIPT, STYLE
-
ISINDEX, BASE, META, LINK (empty)
- OBJECT (see content model below)
-
BODY
- INS, DEL (special rules apply)
- flow
-
block
-
P, H1, H2, H3, H4, H5, H6
-
UL, OL
-
DIR, MENU
-
DL
-
PRE
-
inline excluding
IMG, OBJECT, APPLET, BIG, SMALL, SUB, SUP, FONT, BASEFONT
-
DIV, CENTER, BLOCKQUOTE, IFRAME
-
NOSCRIPT
-
NOFRAMES
-
FORM
-
flow excluding an enclosed FORM
-
ISINDEX,
HR (empty)
-
TABLE
-
CAPTION
- COLGROUP
- COL (empty)
- THEAD, TBODY, TBODY
-
ADDRESS
-
FIELDSET
- inline
-
#PCDATA
-
TT, I, B, U, S, STRIKE, BIG, SMALL, FONT,
EM, STRONG, DFN, CODE, SAMP, KBD, VAR, CITE, ABBR, ACRONYM,
SUB, SUP, Q, SPAN, BDO
-
A
-
inline excluding an enclosed A element
- OBJECT, APPLET
-
IMG, BASEFONT, BR (empty)
-
SCRIPT
-
MAP
-
INPUT (empty)
-
SELECT
-
TEXTAREA
- LABEL
- LABEL excluding enclosed LABEL
- BUTTON
- flow excluding A, INPUT, SELECT, TEXTAREA, LABEL, BUTTON,
FORM, ISINDEX, FIELDSET, IFRAME
Nesting rules for HTML 4.01 Frameset
HTML
- HEAD (content model as above)
- FRAMESET
- FRAMESET (note recursion)
- FRAME (empty)
- NOFRAMES
-
BODY (see content model above) excluding NOFRAMES
In HTML 4.01 Frameset, the content model for NOFRAMES
applies inside the BODY too, instead of the content model for NOFRAMES
in HTML 4.01 Transitional.
The information here is based on the
DTDs, basically the
transitional DTD,
in the
1999-12-24 version
of the
HTML 4.01 Specification.
Note that
XHTML 1.0 is, as its subtitle says,
"A Reformulation of HTML 4 in XML 1.0", so the nesting rules
are the same as in HTML 4.01. However, there are the following
differences that affect the nesting rules:
- The content
of
script
and style
elements is
CDATA in HTML 4 but #PCDATA in XHTML.
- In XHTML, a
table
element may have a tr
element as its direct constituent. In HTML 4.01, that's not
allowed, but note that since the start and end tags of a tbody
element are omissible in HTML 4.01, this is not a big difference.
However, note that when a table
element directly contains
a tr
element, an intervening tbody
element
is implied by HTML 4.01 rules but not
by XHTML rules, and this matters e.g. when you have a style sheet
which uses tbody
as a selector.
Moreover, some of the
restrictions on nesting are expressed differently;
due to metalanguage differences, some limitations are described in prose
only in the XHTML specification, and this implies that a
validator will not catch such violations of the limitations
when validating against XHTML DOCTYPE but will catch them when an
HTML DOCTYPE is used.
See section
Differences with HTML 4
in the XHTML 1.0 Specification.
Date of last update: 2001-04-06
Interested in related documents? See
a list of documents about WWW
written or recommended by me.
Specifically, this document has a sister:
Allowed nesting of elements in HTML 4 Strict (and XHTML 1.0 Strict),
which describes the rules for the Strict versions,
which
include all elements and attributes that have not been deprecated or do not appear in frameset documents.
Jukka Korpela.