Span and div

In HTML, the   and   elements are used where parts of a document cannot be semantically described by other HTML elements.

Most HTML elements carry semantic meaning – i.e. the element describes, and can be made to function according to, the type of data contained within. For example, a  element should contain a paragraph of text, while an   element should contain the highest-level header of the page; user agents should distinguish them accordingly. However, as  and   have no innate semantic meaning besides the logical grouping of the content, they can be used to specify non-standard presentation or behaviour without superfluous semantic meaning.

Differences and default behavior
There is one difference between  and. In standard HTML, a  is a block-level element whereas a   is an inline element. The  block visually isolates a section of a document on the page, in the same way as a paragraph. The  element contains a piece of information inline with the surrounding text. In practice, even this feature can be changed by the use of Cascading Style Sheets (CSS).

Practical usage
and  elements are used purely to imply a logical grouping of  enclosed elements.

When they are labeled with  or   attributes,   and   elements can denote types of information otherwise indescribable with HTML. For example,  may be used to indicate the author's name in a document, and   may be used specifically to indicate a date.

There are three main reasons to use  and   tags with   or   attributes:

Styling with CSS
Perhaps the most common use of  and   elements is to carry   or   attributes in conjunction with CSS to apply layout, typographic, color, and other presentation attributes to parts of the content. CSS does not just apply to visual styling: when spoken out loud by a voice browser, CSS styling can affect speech-rate, stress, richness and even position within a stereophonic image.

For these reasons, and for compatibility with the concepts of the semantic web, discussed below, attributes attached to elements within any HTML should describe their semantic purpose, rather than merely their intended display properties in one particular medium. For example,  is semantically meaningless, whereas   is much more useful. By the correct use of CSS, on the screen 'warnings' may be rendered in a red, small font, but when printed out, they may be omitted, as by then it is too late to do anything about them. Perhaps when spoken they should be given extra stress, and a small reduction in speech-rate. The second example is semantic markup, rather than merely presentational, but serves both purposes when combined with CSS.

Semantic clarity
This kind of grouping and labeling of parts of the page content might be introduced purely to make the page more semantically meaningful in general terms. It is impossible to say how and in what ways the World Wide Web will develop in years and decades to come. Web pages designed today may still be in use when information systems that we cannot yet imagine are trawling, processing, and classifying the web. Even today's search engines such as Google and others use proprietary information processing algorithms of considerable complexity.

The World Wide Web Consortium (W3C) has for some years been running a major Semantic Web project designed to make the whole web increasingly useful and meaningful to today's and the future's information systems.

During the page design process, the designer has a clear idea of exactly the purpose and meaning of each element and sub-element of the content. If possible, and if standard HTML elements exist that express that meaning, they should be used. If not, there is no better time to encapsulate the meaning and purpose in a  or   element with appropriate   or   attributes. If nothing more, doing so will help future editors to maintain the markup.

The microformats movement is an attempt to build on this idea of semantic es. For example, microformats-aware software might automatically find an element like  and allow for automatic dialing of the telephone number.

Access from code
Once the HTML or XHTML markup is delivered to a page-visitor's client browser, there is a chance that client-side code will need to navigate the internal structure (or Document Object Model) of the web page. The most common reason for this is that the page is delivered with client-side JavaScript that will produce on-going dynamic behavior after the page is rendered. For example, if rolling the mouse over a 'Buy now' link is meant to make the price, elsewhere on the page, become emphasized, JavaScript code can do this, but JavaScript needs to identify the price element, wherever it is in the markup, in order to affect it. The following markup would suffice:. Another example is the Ajax programming technique, where, for example, clicking a hypertext link may cause JavaScript code to retrieve the text for a new price quotation to display in place of the current one within the page, without re-loading the whole page. When the new text arrives back from the server, the JavaScript must identify the exact region on the page to replace with the new information.

Less common, but just as important examples of code gaining access to final web pages, and having to use  and   elements'   or   attributes to navigate within the page include the use of automatic testing tools. On dynamically generated HTML, this may include the use of automatic page testing tools such as HttpUnit, a member of the xUnit family, and load or stress testing tools such as JMeter when applied to form-driven web sites.

Overuse
The judicious use of  and   is a vital part of HTML and XHTML markup. However, the overuse of these elements, sometimes called divitis (a common mistake of not only beginners), is itself a minor form of tag soup.

For example, when structurally and semantically a series of items need an outer, containing element and then further containers for each item, then there are various list structures available in HTML, one of which may be preferable to a homemade mixture of  and   elements.

For example, this...

...is usually preferable to this:

Other examples of the semantic use of HTML rather than  and   elements include the use of   elements to divide up a web form, the use of   elements to identify such divisions and the use of   to identify form   elements rather than ,   or   elements used for such purposes.

HTML5 introduces many new semantic elements and attributes that are intended to be used in these ways. A few examples include the,  ,   and   elements.