Archive: mothballed errors

These are examples of errors in HTML that used to be treated differently by different browsers, but are all handled the same these days.

Don't take that as a signal that it's OK to have these errors in web pages though...

Structural error: missing > on start tag

<div
<hr>
... content ...
</div>

Most modern browsers tend to react the same: skip the HR and show only the rest of the content of the div. However, until recently, behaviour was more varied.

Unrecognised tag name with a style (as attribute, or in a stylesheet)

<style> green {color:green} </style>
...
<green style="color:green">this is a green</green>

Table in P

Note: this isn't technically an error, since the p end tag is optional and paragraphs can't contain tables, the p is officially supposed to end on encountering the table.

<p>Before the table
<table><tr><td>in the table</td></tr></table>
after the table

All modern browsers handle this correctly. However, some older browsers, like IE7, did not end the paragraph, but put the table in it instead. So it's best to avoid this situation.

Staggered tag mismatch: span partially in p, partially outside

(In this example, the p is red and the span is underlined, so you can see where each ends.)

Group of radio buttons with more than one checked.

<input type="radio" id="r1" name="group" value="1" checked="checked">
<input type="radio" id="r2" name="group" value="2" checked="checked">
<input type="radio" id="r3" name="group" value="3" checked="checked">

All modern browsers handle this the same: only the last button has a selection mark. But it's best to avoid this situation; there have been browsers in the past, for example Konqueror 3, that showed selection marks on all three buttons.

Note: if none has "checked", one is supposed to be checked automatically, but no browser does this.

Missing end quote after attribute

<div style="color:red>The first div</div>
<div>The second div</div>

Text in a list between two list items

<ul>
<li>The first list item</li>
The text in between
<li>The second list item</li>
</ul>

These days, all browsers act the same: display the text between the li elements. But IE7 seems to ignore all </li> end tags, and it puts the text to the end of the first li (on the same line).

Empty width attribute

Bad link, variant 2: bad file type for the css

Of the browsers that don't use the stylesheet, surprisingly they all do use it in quirks mode.

Bad link 3: encoding mismatch

If the HTML file is encoded as UTF-8, and the CSS file is UTF-16, and the link doesn't explicitly specify an encoding, this happens:

Font names with quotes, without quotes

If you have a font name that contains a quote, or an apostrophe, you have to be extra careful how to use it.

(And yes, this is a real font; it's in the Linux repositories.)

Modern browsers all consider the last one OK, but not the first two.

<![CDATA[ sections in HTML

<div>CDATA sections in HTML are <![CDATA[not]]> invisible.</div>

In modern browsers, the CDATA section is not shown.

getElementById without ID

<a name="lime">A lime a</a>
<script type="text/javascript">
document.getElementById('lime').style.backgroundColor = 'lime';
</script>

The <a> does not have an ID, so its background colour is not changed in modern browsers.

Bad value for align attribute

Note: as align is deprecated, this test was done using HTML 4.01 Transitional.

With XHTML: ID in the XML namespace

All elements and attributes are normally taken from the HTML namespace, but it's possible to use other namespaces by using a prefix. So in this example, the id on the bottom div is in the XML namespace; it won't have any effect on how the div is handled.

<style type="text/css"> #theID {color:green} </style>
.
.
<div id="theID">This is a div with an ID</div>
<div xml:id="theID">This is a div with an xml:ID</div>

Most browsers ignore the id in the bottom div and won't give it the style. Except Opera v12 and below, which will.

It has come to my attention however, that xml:id does not exists in the formal recommendations; that it was only part of a proposal in the past. So apparently, Opera already anticipated the future, but it never materialized.