W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > July to September 1998

RE: The problem with tables

From: David Bolnick <davebo@MICROSOFT.com>
Date: Fri, 24 Jul 1998 12:57:25 -0700
Message-ID: <FFD1BA74C6A7D111A09500805F9F88F502E7A6C5@red-msg-43.dns.microsoft.com>
To: WAI <w3c-wai-ig@w3.org>
This is strictly my opinion - or maybe it's a question:

Since HTML 3.2 there has been a good way to distinguish a columnar table
from a table of rows (by use of COL and COLGROUP). Maybe we should encourage
authors to employ this method and screen reader ISV's to use this method as
a way to resolve HTML tables. I have attached a quick-and-dirty example that
is based on these HTML 3.2 tags. 

BY ROWS:

<TABLE>
	<Caption>This is a table organized by rows</Caption>
	<TR>
	<TD>Row1, Col1
	<TD>Row1, Col2
	<TR>
		<TD>Row2, Col1
		<TD>Row2, Col2
</TABLE>

BY COLUMNS:

<TABLE>
	<Caption>This is a table organized by columns</Caption>
<COLGROUP>
	<COL Align=right>
	<COL Align=left>
<COLGROUP>
	<COL Align=center>
<TBODY>
	<TR>
	<TD>Column Group1, Row1, Col1 (right aligned)
	<TD>Column Group1, Row1, Col2 (left aligned)
	<TD>Column Group 2, Row1, Col3 (centered)
	<TR>
	<TD>Column Group1, Row2, Col1 (right aligned)
	<TD>Column Group1, Row2, Col2 (left aligned)
	<TD>Column Group 2, Row2, Col3 (centered)
</TABLE>
	________________________
	David A. Bolnick
	Accessibility Program Manager: Multimedia, Telecommunications
	Microsoft Corporation, One Microsoft Way, Redmond, WA  98052
	E-mail:  <mailto:davebo@microsoft.com> davebo@microsoft.com     Web:
<http://microsoft.com/enable> http://microsoft.com/enable


		 <<Row_Col.htm>> 

		-----Original Message-----
		From:	Charles McCathieNevile
[mailto:charlesn@sunrise.srl.rmit.edu.au]
		Sent:	Thursday, July 23, 1998 11:52 PM
		To:	WAI
		Subject:	The problem with tables

		TABLE is probably the most misused element in the HTML spec.

		Using tables to provide columns (for examples see Microsoft,
W3C, News 
		services, or most other people who have spent a bit of money
on web 
		design) causes significant problems for screen readers.

		(The following is an attempt to demonstrate what happens. To
mimic a 
		screen reader, read as though it were ordinary text, left to
right all 
		the way across the page)

		(Start example)
		              For Lynx Users there is For users of screen
readers the
		home page     no real problem, since  problem is that they
cannot tell
		              it ignores table tags,  where there is a
column. Instead
		Search this   effectively, but quite  the reader just reads
the lines
		Website       crudely. Unfortunately  all the way across the
screen.
		              not all people can use  This makes it so
difficult to
		Look at me    Lynx, for any number of understand what is
being read.
		              good reasons.
		<end example)

		If you read that example all the way across, it sounds like
a lot of 
		nonsense. It is in fact two well formed paragraphs, and some
short labels 
		on the side. Many Screen readers do exactly the same with
text in columns.

		The amount of effort required to put text in columns (or
images with 
		appropriate ALT text) compared with a simpler more
comprehensible layout 
		does not seem to justify erecting this particular
accessibility barrier. 
		Add this to the fact that it is a blatant violation of the
principles 
		explained in the HTML specification, that form and content
ought to be 
		seperate.

		A proper use of TABLE is to present tabulated data, rather
than relying 
		on PRE to do it. This means that it is possible to linearise
the table, 
		and reformat the data without losing the meta-information
provided by the 
		table structure. (It still causes problems for screen
readers, but at 
		least there is something to say in its favour)

		An improper use of tables, but one which, so far as I know
has very 
		limited accessibility problems, would be to format material
in such a way 
		that in each of two columns images were alternated with
text. Then only 
		the ALT text of the image could give rise to a
misunderstanding, a 
		situation which could be handled by an intelligent author.

		A really stupid use of tables can be found  on most
commercial websites, 
		where they are used to cram so much information into the
visible window 
		of a graphics browser that the site becomes very difficult
to understand 
		for sighted people who have a high-quality connection and a
good monitor.

		While CSS is not implemented, using it to control
presentation does not 
		cause major problems. Perhaps when it is implemented, and
users have the 
		ability to over-ride a style sheet provided by an author
which does not 
		suit their particular needs, the problem for screen readers
will have 
		been effectively solved. Otherwise, the whole text in
columns debate will 
		rise again. But there are two ways around it.

		The first is to make use of frames. It is perfectly possible
to produce a 
		site which is based on the use of frames, and which is
accessible. It's 
		just that I am unaware of any such site currently in
operation. (But then 
		most websites are inaccessible - it's not peculiar to
frames.) It does 
		take a small amount of extra work, but bad design is
generally not harder 
		than good design, it just requires thinking along different
principles.

		The second way is to ensure that there is a non-columnar
version of 
		everything in columns, and that it is clear from the start
of the 
		columnar data where to find it. This can take a considerable
amount of work.

		The alternative is to abandon the wholesale misuse of
tables. My Doctor 
		tells me not to smoke or live on a diet of beer, fish and
chips and 
		steak. My mechanic tells me not to drive my car as fast as
it goes all 
		the time. My accountant tells me not to spend all my money
on beer, fish 
		and chips, steak and car repairs.

		As experts in their field, I listen to their advice.
Otherwise I'd save 
		myself the money. If a doctor told me beer, tobacco and fat
were going to 
		make me healthy, I would walk away very quickly. Why then,
as experts 
		ourselves, are we so scared to tell people what we know?

		Isn't the American government suing tobacco companies for
withholding 
		information? Apart from the fact that our product doesn't
kill many 
		people, what makes us better than those companies?

		Ignorance could be used as an excuse, but the reason for
fora like this 
		is to end our ignorance by sharing our knowledge. And to
encourage others 
		to do the same.

		(end rant. must be friday afternoon)

		Charles McCN
		just my 2c worth

begin 600 Row_Col.htm
M/$A434P^#0H\0D]$63X-"@T*0ED@4D]74SH-"@T*/%1!0DQ%(')U;&5S/6%L
M;#X-"@D\0V%P=&EO;B!V86QI9VX]8F]T=&]M/E1H:7,@:7,@82!T86)L92!O
M<F=A;FEZ960@8GD@<F]W<SPO0V%P=&EO;CX-"@D\5%(@8F=C;VQO<CTB(T%$
M1D8R1B(^#0H)"3Q41"!7:61T:#TQ,#`^4F]W,2P@0V]L,0T*"0D\5$0@5VED
M=&@],3`P/E)O=S$L($-O;#(-"@D)/%1$(%=I9'1H/3$P,#Y2;W<Q+"!#;VPS
M#0H)/%12(&)G8V]L;W(](B-&1D0W,#`B/@T*"0D\5$0^4F]W,BP@0V]L,0T*
M"0D\5$0^4F]W,BP@0V]L,@T*"0D\5$0^4F]W,BP@0V]L,PT*/"]404),13X-
M"CQ"4CX-"CQ(4CX-"CQ"4CX-"D)9($-/3%5-3E,Z#0H-"CQ404),12!R=6QE
M<SUA;&P^#0H)/$-A<'1I;VX@=F%L:6=N/6)O='1O;3Y4:&ES(&ES(&$@=&%B
M;&4@;W)G86YI>F5D(&)Y(&-O;'5M;G,\+T-A<'1I;VX^#0H\0T],1U)/55`^
M#0H)/$-/3"!!;&EG;CUR:6=H="`@8F=C;VQO<CTB(T%$1D8R1B(^#0H)/$-/
M3"!!;&EG;CUL969T("`@8F=C;VQO<CTB(T9&1#<P,"(^#0H\0T],1U)/55`^
M#0H)/$-/3"!!;&EG;CUC96YT97(@8F=C;VQO<CTB(T0S1#-$,R(^#0H\5$)/
M1%D^#0H)/%12/@T*"0D\5$0@5VED=&@],C$P/D=R<#$L(%)O=S$L($-O;#$@
M*')T(&%L:6=N960I#0H)"3Q41"!7:61T:#TR,3`^1W)P,2P@4F]W,2P@0V]L
M,B`H;'0@86QI9VYE9"D-"@D)/%1$(%=I9'1H/3(Q,#Y'<G`R+"!2;W<Q+"!#
M;VPS("AC="!A;&EG;F5D*0T*"3Q44CX-"@D)/%1$/D=R<#$L(%)O=S(L($-O
M;#$@*')T(&%L:6=N960I#0H)"3Q41#Y'<G`Q+"!2;W<R+"!#;VPR("AL="!A
M;&EG;F5D*0T*"0D\5$0^1W)P,BP@4F]W,BP@0V]L,R`H8W0@86QI9VYE9"D-
="CPO5$%"3$4^#0H-"CPO0D]$63X-"CPO2%1-3#X=
`
end
Received on Friday, 24 July 1998 15:57:04 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 19 July 2011 18:13:40 GMT