The Semantic Web has been discussed and debated by academics for years, and really, we are not getting anywhere fast. This post will discuss how to start building semantics into your website from the ground up. While doing this won't create the create the great Semantic Web straight away, it will provide an underground layer of accessible data which applications can start processing and using.

Semantic Tagging

I wrote a brief guide on Semantic Markup so that would be a good place to start. The main idea behind this is to start putting meaning into your web documents.

I imagine most sites in three layers: the content, the presentation and the behaviour. The behaviour represents client side scripting e.g. Javascript, the presentation layer is provided by CSS. The content layer should be clean from anything except markup, if this markup is semantic it means that any application can get a good understanding of the makeup of the document e.g. An application could extract code, make an index or a screen reader could put emphasis on words marked up as such.

While this won't add any value to the Semantic Web it provides a structured base from where everything else can be built on.

Microformats

Microformats are little pieces of structured markup that convey a certain piece of information like a content card or an event or an address. At the moment I have a content card on my site and it gets picked up by the microformats toolbar I have installed, this means that if anyone else with the toolbar comes across my site they can download my contact card if they wish.

Microformats also provide the basis for the Data Web, which I will come to a bit later.

XML

Ok, so we have a bed of semantic markup with some microformats thrown in. Now what?? Well, now we can have some fun. XML can be inserted into any XHTML document via namespaces, however this isn't to say all browsers support it yet (Internet Explorer struggles with this, Firefox works great with it).

What you can do is insert other languages into your site. Have some math to put into your website? Use MathML. Have some graphs you want to display? Use SVG.

I am currently working on a markup language that when combined with an XSLT document creates graphs out of datatables (think normal HTML tables), it is still in an early stage of development though(only bar charts at the moment and some basic document formatting).

RSS and the Semantic Web

So here we are at RSS, this can basically be used to tie applications with the web. The application could download the feed, find the relavent microformats and put them in the contact book, or the calendar. You can probably already see the implications for this.

This is my idea of the Data Web, unlike the Semantic Web, the sites are not linked via RDF, at the moment each site is its own island and while computer applications and websites with access to feeds can take advantage and use the data there is no mass meaning set up yet.

The Semantic Web will come about when sites start producing RDF documents describing the relationships between different URIs. However I believe without an underlying framework of content that has been marked up correctly it makes the entire process a bit useless.