# Specification file for the subset of HTML to be permitted in RFCs (DRAFT)
#
# $Id: permitted,v 1.2 1997/07/28 14:34:41 esr Exp $
#
# The theory is that rendering the HTML to ASCII via lynx -dump has to
# not lose information.
#
# To generate this set, we started with HTML 3.2.  First, we threw out every
# form of interactive markup (applets, forms, isindex, etc).  Then we tossed
# out IMG and AREA tags, so that every RFC will remain just one document
# without a requirement for auxilliary images to be hauled around.  Then we
# chucked the deprecated tags LISTING and XMP. 
#
# Then we threw out the physical style tags B, I, TT, U, FONT, BIG, SMALL,
# and STRIKE to enforce strctural markup.
#
# The BODY attributes ALINK BGCOLOR LINK TEXT VLINK that affect HTML
# presentation only over the whole document are theoretically OK, but
# we exclude them so HTML renderings will have an unfussy, uniform look.
#
# We have retained all the structure tags CITE, CODE, DFN, EM, KBD, SAMP,
# STRONG and VAR in spite of the fact that lynx -dump does not render them
# all as distinct (CITE, EM, and STRONG are rendered with underline; CODE, DFN,
# KBD, SAMP, and VAR have no effects).  This is to assist in Postscript 
# rendering.  Authors must be warned that changes between indistinguishable
# fonts may not be used to convey information.
#
# Table markup ought to be permitted, and will be as soon as lynx or any
# other HTML-to-text converter is able to handle tables.  (Hmmm...it's
# been pointed out that tables are problematic for non-visual rendering.)
#
# Allow the following HTML tags and attributes.  These are a subset of 3.2:
#
A		HREF NAME REL REV TITLE
ADDRESS
BLOCKQUOTE
BODY
BR		CLEAR
CENTER
CITE					# Lynx renders as underline
CODE					# No effect under Lynx
DD
DFN					# No effect under Lynx
DIR
DIV		ALIGN
DL		COMPACT
DT
EM					# Lynx renders as underline
H1		ALIGN
H2		ALIGN
H3		ALIGN
H4		ALIGN
H5		ALIGN
H6		ALIGN
HEAD
HR		ALIGN NOSHADE WIDTH
HTML		VERSION
KBD					# No effect under Lynx
LI
LINK		HREF REL REV TITLE
META		CONTENT CHARSET NAME	# CHARSET?
OL		COMPACT START TYPE
P		ALIGN
PLAINTEXT
PRE		WIDTH
SAMP					# No effect under Lynx
STRONG					# Lynx renders as underline
TITLE
UL		COMPACT STYLE
VAR					# No effect under Lynx
#
# The caption/table/td/th/tr elements are a problem.  In theory they
# should be fine, but in practice I don't know of any translator that
# can compile them to text
#
#CAPTION		ALIGN VALIGN
#TABLE		ALIGN BORDER CELLPADDING CELLSPACING WIDTH
#TD		ALIGN COLSPAN HEIGHT NOWRAP ROWSPAN VALIGN WIDTH
#TH		ALIGN COLSPAN HEIGHT NOWRAP ROWSPAN VALIGN WIDTH
#TR		ALIGN VALIGN
#
# Allow HTML 3.2 entities that have printable rendering in 7-bit ASCII
#
"	# quotation mark
&	# ampersand
<	# less than
>	# greater than
 	# only proposed in 3.2, not officially standard