Re: Ampersands + "Last modified:" Malfunctions

At 12:39 03/06/00 -0400, Alan C. Baird wrote:
>Hi!
>
>I'm not a member of this list, but I just wanted to note a couple of 
>validator bugs... at:
>
>  http://www.htmlhelp.com/tools/validator/problems.html#amp
>
>, the documentation suggests:
>
>> To avoid problems with both validators and browsers, always use 
>> & in place of &:
>>
>> <a href="foo.cgi?chapter=1&amp;section=2">...</a>
>
>However, many CGI engines will not accept &amp; - and rather than 
>forcing all those webmasters out there to stop programming in the 
>commonly-accepted fashion, why not change the validator to accept 
>ampersands, when they're embedded in anchor tags?  Half my pages 
>won't validate correctly, because I've linked to CGI sites which 
>operate only with ampersands.  It seems like we've got the cart 
>before the horse, here.  ;-)

I had this gripe recently.

Annoyingly the people who wrote the cgi spec were slightly ignorant of SGML.
Consequently we have two incompatible specs and they cannot be made to meet 
beautifully. 

The answer I was given on this list was to code the & symbol as &amp; in URLS, 
it is generally claimed that this will work with all cgi scripts (this may or may not be true).

My winge was that there is no way within SGML to specify that 
attribute values are CDATA, they have to be PCDATA, which means that the 
& character is magic. 

Have you not been caught by a CGI script which takes a parameter called section or para?

You will find, to your horror, due to a bug in Netscape (and mysteriously IE), 
that you do not need to place the final semicolon in the entity reference so 

http://www.asite.com/cgi-bin/t.pl?somparam=45&para=34&section=987 is 
rendered into para and section symbols :(

yours
peeved
timp


Member of http://www.paneris.org/

Received on Tuesday, 20 June 2000 14:22:41 UTC