Re: Strange definition of Frame in Cougar DTD

Arnaud Le Hors (lehors@w3.org)
Tue, 29 Apr 1997 17:29:25 +0200


Message-ID: <336613D5.794E@w3.org>
Date: Tue, 29 Apr 1997 17:29:25 +0200
From: Arnaud Le Hors <lehors@w3.org>
To: Masayasu Ishikawa <mimasa@vega.aichi-u.ac.jp>
CC: www-html@w3.org
Subject: Re: Strange definition of Frame in Cougar DTD

Masayasu Ishikawa wrote:
> 
> Hi,
> 
> I checked 18-Apr-97 version of Cougar DTD and found a strange definition
> related to Frame. In Cougar DTD, FRAMESET element is defined as:
> 
>         <!ELEMENT FRAMESET - - (FRAMESET|FRAME)+>
> 
> There's no space for NOFRAMES, isn't it? It should be
> 
>         <!ELEMENT FRAMESET - - (FRAMESET|FRAME|NOFRAMES)+ >
> 
> or at least
> 
>         <!ELEMENT FRAMESET - - ( (FRAMESET|FRAME)+ & (NOFRAMES?) )>
> 
> I think NOFRAMES element should be defined as COMPULSORY element
> for Frame document, because without NOFRAMES, Frame document becomes
> totally useless for user agents that don't support Frame.

This is not strange, this is in agreement with the latest Working Draft
on Frames [1]. The answer to your request is at the end of the DTD:

<!ENTITY % html.content "HEAD, FRAMESET?, BODY">

Which means authors can provide a whole BODY for user agents that don't
support frames. In addition, NOFRAMES can be used inside BODY for
content to be only rendered by user agents that do not support (or
simply do not use) frames, giving even more control.

[1] http://www.w3.org/pub/WWW/TR/WD-frames-970331
-- 
Arnaud Le Hors - W3C, User Interface Domain - http://www.w3.org