"&" and valid html

**blackdog** · May 17, 2004, 19:07

I have a phone reviews site and I have AT&T. When I try to validate my site, the "&" causes all sorts of errors.

For example...

<LI>Line 151, column 52: entity was defined here

<td class="navnorm"><a class="navtop1" href="at&t.php">AT&T</a></td>

<LI>Line 151, column 63: cannot generate system identifier for general entity "T"

...class="navnorm"><a class="navtop1" href="at&t.php">AT&T</a></td>

anyone know how to fix this so my site validates? thanks!

**KLB** · May 17, 2004, 19:13

Simply use & any time you want &.

**Imminent** · May 17, 2004, 19:15

You really shouldn't use ampersand in your file names.

Change it to at-and-t.html or atandt.html

**blackdog** · May 17, 2004, 19:19

thanks for the advice guys!

**KLB** · May 17, 2004, 19:21

Originally Posted by Imminent

You really shouldn't use ampersand in your file names.

Change it to at-and-t.html or atandt.html

Very good point! I simply noticed the question about ampersand and did not notice it was being used in a URL. & should only be used in URL query strings (e.g. bla.html?variable1=yes&variable2=no)

**cfm** · May 18, 2004, 03:11

Off Topic:

I seem to recall that there's a property you can set in .htaccess in Apache that will allow you to specify & instead of & for query string variables. I forget what it is specifically, and I'm not sure if there's one for IIS, but I thought I'd mention that.

**KLB** · May 18, 2004, 13:45

Originally Posted by compuwhiz7

Off Topic:

I seem to recall that there's a property you can set in .htaccess in Apache that will allow you to specify & instead of & for query string variables. I forget what it is specifically, and I'm not sure if there's one for IIS, but I thought I'd mention that.

?????

If you use & instead of & in the querystrings of links, the webbrowser will translate & into & and only submit & when a link is clicked on. There is no need for a special configuration Apache or IIS.

**m2k** · May 18, 2004, 13:48

Also, an example:

you create a site with valid XHTML and CSS and want to include links at the bottom of the pages for each. If you copy and paste the url's (generated by the W3C site) of your validated pages into your links, don't forget that the url's have "&"s in them. Go through and change them to &amp's or, technically, you will wind up with invalid XHTML despite just validating it (if that makes sense).

**cfm** · May 18, 2004, 13:52

Originally Posted by KLB

?????

If you use & instead of & in the querystrings of links, the webbrowser will translate & into & and only submit & when a link is clicked on. There is no need for a special configuration Apache or IIS.

Yep, but I do seem to recall that you can specify what delimiter to use. I honestly can't remember what the application was, but I do remember the existence of such a setting.

**Rincewind** · May 18, 2004, 17:28

The & think in the w3c validator is a bug in the validation system. Though the & throughs an error, it's no actually invalid code. It's just that the vallidator script is not smart enough to work it out.

The w3c validator script is a tool to help you produce valid code. It is not the final word on what is actually valid or not. That is left to the various specifications and guidlines. When using a vallidator, it is more important to understand why your code fails and be able to make an informed decition based on the results than it is to get a clean sheet.

**KLB** · May 18, 2004, 18:00

Originally Posted by Rincewind

The & think in the w3c validator is a bug in the validation system. Though the & throughs an error, it's no actually invalid code. It's just that the vallidator script is not smart enough to work it out.

Using the & sign directly attached to a string of characters when & should be used is actually a validation error and not a bug with validators. The reason is that the ampere is a modifier that lets the browser know that the next characters are a code for a specific ASCII character. Originally (circa Netscape 1 and/or 2) if one used the '&' sign, the browser would try to convert the attached string of characters into an ASCII character, which frequently caused problems. Because people had a bad habit of not using '&' correctly, browsers were modified to compensate just as we have seen browsers modified to deal with other bad coding habits like not closing a table.

**Rincewind** · May 18, 2004, 20:26

Using & in your plain text is an error since & is the correct method. However, inside url's or as I often find inside my javascript code, the conventions of html do not and should not apply. The file naming conventions are determined by the operating system of you server not the w3c. If the OS allows AT&T.php as a file name the the URL to that file should be AT&T.php not AT&T.php. Your browser corrects the error of the & in the url and sends AT&T.php to the server. If you actually tried bypassing the browser (e.g. use an ftp, ssh or telnet client) and asked the server for the file AT&T.php then you'd get a 404 error. Thus using AT&T.php is incorrect file name.

**Hulkur** · May 19, 2004, 04:50

in php.ini

Code:

arg_separator.output = "&amp;"

or .htaccess

Code:

php_value arg_separator.output &amp;

lets you specify wether to use & or & (for automatic attaching of session ID to url)

**fatnewt** · May 19, 2004, 08:49

Originally Posted by Rincewind

Using & in your plain text is an error since & is the correct method. However, inside url's or as I often find inside my javascript code, the conventions of html do not and should not apply. The file naming conventions are determined by the operating system of you server not the w3c. If the OS allows AT&T.php as a file name the the URL to that file should be AT&T.php not AT&T.php. Your browser corrects the error of the & in the url and sends AT&T.php to the server. If you actually tried bypassing the browser (e.g. use an ftp, ssh or telnet client) and asked the server for the file AT&T.php then you'd get a 404 error. Thus using AT&T.php is incorrect file name.

Uhm, this isn't true.

True, & is not a part of the actual URL. But it is the correct way to reference the ampersand character in (X)HTML. Since you need to reference this character in the URL, you use & (because even though this is a URL, it's a part of an (X)HTML document). The browser knows this, so it translates it for you when it actually goes to this page.

Here it is in the XHTML recommendation:

Originally Posted by W3C on XHTML

For example, when the href attribute of the a element refers to a CGI script that takes parameters, it must be expressed as http://my.site.dom/cgi-bin/myscript.pl?class=guest&name=user rather than as name=userhttp://my.site.dom/cgi-bin/myscript.pl?class=guest&name=usr.

This becomes especially important if your XHTML document is being read by an XML interpereter, which is why it's mentioned here... but the same applies for HTML.

The SitePoint Forums have moved.

User Tag List

Thread: "&" and valid html

Thread Tools

Display

"&" and valid html

Bookmarks

Bookmarks

Posting Permissions