Semantics Is The New Black

30 Jul

Every year around January time, the design/development community make a few predictions as to what will be the big thing for the upcoming year. Predictions range from popular colours, site types, font choices as well as more esoteric things such as concepts (AJAX was touted as the coming thing this year with some apparent justification) but a few things become popular due to events or industry leaders making them news (for example Andy Clarke’s recent post about accessibility and societal control and SiteMorses recent footshooting debacle has placed accessibility back to the forefront of the community’s collective mind).

And then some things quietly and unobtrusively instill themselves into our design/development lives with scarcely a ripple.

The ongoing movement towards semantics on the web is something that does seem to pass by even us in the community responsible for its promotion. I want to take a look at a few things that we might not even have thought of as examples of web based semantics and how they are affecting us on a daily basis.

What we mean by semantics as they apply to the web is the principle of the ‘thing’ itself having meaning as well as the message that the ‘thing’ is overtly conveying. A prime example of this:

This is a paragraph.

You can’t get much more semantic than that! We use a ‘paragraph’ element to convey the covert meaning on the section in question as well as to display the text in that element overtly.

But these days, semantics cover a much wider range of possibilities and meanings than a simple markup element. Lets take a look at Search.

Search engines such as Google, Yahoo and MSN will place an increasing amount of importance on semantics. This process is already underway – I’ve discussed before how Google are implementing a process called Latent Semantic Indexing – and will only increase pace. But what does semantics mean for search engines? It can mean lots of things. Firstly there is the semantic relationship between the search word/phrase you use to generate results and the actual results themselves. Obviously, the better that match is the more accurate your SERPs (Search Engine Results Pages) will be.

From a web developers point of view, semantics affect our sites relationship with search engines in two major ways. Firstly if you want to promote the phrase ‘bad credit loans’ on your site then creating phrases that share a semantic meaning with that phrase or the words in that phrase is a good idea ‘bad credit loans’ could be semantically matched with ‘debt consolidation’ or ‘secured loans’ or ‘credit worries’. The second way semantics is important to us comes in terms of the sites that link to us. If I’m in charge of a travel insurance website then my automatic assumption might be to get lots of backlinks from finance related sites. However, the semantic way of looking at the relationship would be to get links from sites that share a common or similar theme – holiday sites, airline sites etc.

A more intriguing and tantalizing possibility regarding semantics and search engines is the possibility that search engines are capable of determining the _type_ of site. By this I mean is the site an e-commerce site? Is it a forum? Is it a basic brochure site? Is it a blog? This semantic relationship between the underlying code of a site, its structure and its overall purpose does seem detectable by engines albeit in a fairly basic ‘brute force’ way – so far.

Moving away from search a little bit we should take a look at how blogging has powered a massive increase in constructing a semantic structure to its particular environment. Sites like Technorati which are essentially search engines for blogs have a core functionality which lists all the other sites a particular site is receiving links from – in the blogosphere links are awarded by bloggers who feel the linkee shares a common goal/spirit/language/understanding with them and hence Technorati’s Cosmos feature is a foundation of semantics – communication going beyond just the overt. With blogs becoming increasingly popular its no wonder the big search engines are interested in matching sites like Technorati’s semantic influence.

Then of course there are the blogs themselves – categorisable and taggable as sites never have been before and capable of creating a vast community based not just on what each blogger finds interesting but on the way that blogs store, produce and display information. Again, the way its said is as important as whats actually _being_ said. And as new formats and new offshoots appear (del.icio.us/ and flickr for example) that semantic relationship between blogs that share no visual similarities and _who might not even be aware of each other_ builds and builds. Flickr and del.icio.us can be fed into a lot of blogs and blogs can export their content in meaningfully rich ways via RSS.

So, semantics – its the new black. As our understanding of what can be achieved by making sure we write to a common format and how relationships between codable structures fire relationships between people increases so will our ability to have a web that can finally begin to bring things to us with increasing accuracy. The future isn’t Search, the future is Delivery.

3 Responses to “Semantics Is The New Black”

  1. Ian August 1, 2005 at 20:58 #

    The problem I forsee with many of these tag initiatives will appear down the road, in the same way that deficiencies in html, meta-tags etc only came to light for the most part when the web started getting large. I used the web as part of my job back in 93/94 and there was very little on it, so search was less of a problem, plus there was no one trying to direct you to their website for commercial purposes, with all their attendent tricks to get a high rating with the search engine algorithms.

    The semantic web initiative has been a long time in fruition, but standards like OWL are now here – allowing one in theory to define ontologies for many, many disciplines, with their semantics clearly defined (ok it takes a long time and is boring, but so does building a dictionary). If tools were available that could allow one to tag from a predefined set of tags, using standard ontologies, then I think we’d end up with a more useable web in the future. However, the chances of this seem more distant every day, since the impetus seems to be with flikr, delicious, technorati, etc. who’ve made it easy to add tags (as an aside I’ve never been convinced that Cycorp’s Cyc system could be practically leveraged for this problem, even though it seems an obvious match). The big problem with self-tagging comes when my tag and your tag have the same name symbol, but different semantics.

    The one place I see these simple tagging schemes being of most use is inside organisations (e.g. on their intranet) for knowledge management purposes. There, there tends to exist more of a common understanding of terms, and the perceived cost/reward would be favourable.

    Ian

  2. JiggyWittit March 6, 2006 at 11:52 #

    Kewl blog you got goin on up here.
    Peace, JiggyWittit

Comments are closed.

%d bloggers like this: