Re: Characters in attribute values

=?iso-8859-1?Q?Martin_J=2E_D=FCrst?= (mduerst@ifi.unizh.ch)
Mon, 18 Aug 1997 15:31:38 +0200 (MET DST)


Date: Mon, 18 Aug 1997 15:31:38 +0200 (MET DST)
From: =?iso-8859-1?Q?Martin_J=2E_D=FCrst?= <mduerst@ifi.unizh.ch>
To: Holger Wahlen <wahlen@ph-cip.Uni-Koeln.DE>
cc: www-html@w3.org
In-Reply-To: <199708142019.AA07609@jupiter.ph-cip.Uni-Koeln.DE>
Message-ID: <Pine.SUN.3.96.970818153008.675D-100000@enoshima>
Subject: Re: Characters in attribute values

On Thu, 14 Aug 1997, Holger Wahlen wrote:

> After my question about a passage of the SGML declaration in
> the HTML 4 draft ...
> 
> | ----------quote----------
> |     NAMING   LCNMSTRT ""
> |              UCNMSTRT ""
> |              LCNMCHAR ".-"    -- ?include "~/_" for URLs? --
> |              UCNMCHAR ".-"
> | ----------/quote----------

> I'm thinking of values for attributes like HREF, SRC and so
> on. With the present declaration, it's valid to have
> 	<A HREF=file.html>
> instead of
> 	<A HREF="file.html">,
> but the quotation marks can't be left out in the (frequent)
> case that there's a directory change involved:
> 	<A HREF="subdir/index.html">
> has to remain that way. No big deal, but it's quite easy to
> forget the closing quotation mark, especially for longer
> paths.

The problem is that the more you add, the more difficult it
will be to remember which are allowed and which not. URLs
can contain quite a few special characters (but not others),
and so on, and hardly anybody knows them all.

Regards,	Martin.