W3C home > Mailing lists > Public > www-style@w3.org > October 1997

RE: CSS1 and tables

From: Liam Quinn <liam@htmlhelp.com>
Date: Tue, 07 Oct 1997 17:46:05 -0400
Message-Id: <>
To: "Chris Wilson (PSD)" <cwilso@MICROSOFT.com>, www-style@w3.org

At 02:21 PM 07/10/97 -0700, Chris Wilson (PSD) wrote:
>> Liam Quinn [SMTP:liam@htmlhelp.com] wrote:
>>Rather, the bug seems to be caused by IE3 and IE4 allowing a table to
>>be contained within a paragraph.  IINM, this is a long-standing issue
>>with IE, since it would always (and still does) render
>>       <P ALIGN=center><TABLE>...</TABLE>
>>with the table incorrectly centred.
>If we could strictly follow the DTD, this would not happen.  For the
>same backwards-compatibility reasons I've been talking about with table
>cell rendering properties, though, we can't.

What backwards-compatibility?  When you say "backwards-compatibility" I 
take it to mean, in this context, bug-for-bug compatible with Netscape and 
the kind of authoring common on the Web.  Yet Netscape 1.22, 2.02, 3.04, 
and 4.03 all render

	<P ALIGN=center><TABLE>...</TABLE>

with the table correctly left-aligned.  Given this, I would imagine more 
authors depending on a paragraph preceding a TABLE to be closed by the 
<TABLE> start-tag than on the TABLE being centred.

>Hmm.  I don't want to be seen as lending direct credibility to that
>statement - my opinion is that Netscape perpetrated that "original web
>design", and the sole reason we have had to duplicate its unfortunate
>side effects is that in order to increase our user base, we must be
>compatible with the majority of content that's out there.

Does this mean that if MSIE were to takeover from Netscape as the market 
leader we'd see a lean towards proper rendering over popular rendering?  
Netscape has occasionally altered its handling of bad HTML (e.g., 
requiring closing quotes on attributes as of Netscape 2.0, requiring a 
semicolon terminating otherwise ambiguous entities in Netscape 4.0) 
without any effect on its market share, despite some outcries from authors 
who had to fix their documents.  As an author, it would be nice to be able 
to depend on valid HTML not causing rendering problems with a browser.

>A stunning
>percentage of this content was authored in a text editor and "validated"
>by viewing in a browser that believed in this "original web design", and
>therefore may rely on this defaulting of rendering properties inside
>table cells.

Correct me if I'm wrong, but the "default" rendering we're talking about 
is where text-level elements (B, I, FONT) containing a TABLE are generally 
not applied to the contents of the TABLE, but are applied to content 
following the TABLE.  To me, this seems like an unintuitive way to handle 
bad markup, and it seems like it's against what the author desired.  I've 
seen many authors ask why MSIE allows FONT to be applied to an entire 
TABLE but Netscape does not; I remember threads on Usenet where designers 
were contemplating how to get Netscape to change its behaviour, and how to 
change the HTML 3.2 Recommendation at the same time.  Unless I'm 
misinterpreting the behaviour we're talking about, preserving it for 
reasons of legacy is not what the majority of (clueless) authors want.

Version: PGP for Personal Privacy 5.0
Charset: noconv


Liam Quinn
Web Design Group            Enhanced Designs, Web Site Development
http://www.htmlhelp.com/    http://enhanced-designs.com/
Received on Tuesday, 7 October 1997 17:45:31 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:26:45 UTC