When the term 'semantic markup' (or 'semantic html') is used, it generally refers to the use of
descriptive markup as opposed to
presentational markup.
Descriptive (aka. Procedural [
more]) markup includes all those elements and structures that we use to give meaning to the HTML document.
For example:
- Unordered Lists = <ul>
- Definition Lists = <dl>
- Tabular Data = <td>
- Paragraph = <p>
- Address = <address>
- Emphasis = <em>
Presentational (HTML) markup includes items which related to the
presentation (duh) of the document rather than describing its content.
For example:
- Bold = <b>
- Blink = <blink>
- Font = <font>
To finish up, the <marquee> is not to be included as
semantic because it does not give particular meaning to whatever it contains within its opening and closing tags. The purpose of the marquee is purely presentational, as I hope you can understand now. The easiest way (for me) to separate these two types of markup is to imagine (or actually create) a scenario where the HTML document was not displayed in a regular visual browser, but read out by a screen-reader. The semantic elements all work together to create a well structured document that the screen-reader can 'read' out loud. The presentational markup is of no use in this instance (imagine a screen reader announcing blinking text!
) so would go ignored. Lets not write HTML that will be ignored!