Semantic Web for Business Owners
From GetSemantic
Contents |
[edit] Why should I bother putting stuff on the Semantic Web?
For the same reason you put things on the Web - because you want people to use that information. If you publish your phone number on a page, it means that you want someone to be able to get that number and call you - presumably in order to do business. If you are a restaurant, you may publish your menu on your site because you want people to be able to see what you offer.
Last time I (Tom) visited Boston, I wanted a pizza. I cracked open my laptop and searched for a pizza restaurant near my hotel. I found a place, downloaded their menu, had a look and saw what types of pizza they offered and then decided that I'd give them a try. I wasn't disappointed.
The Semantic Web - properly implemented - should allow me even greater control. If I want to have a spicy, vegetarian pizza in Boston, I should be able to fine-tune my search of the city's restaurants to see who offers a pizza with spicy ingredients that is suitable for vegetarians. And I should be able to see their opening times (or, rather, if they are closed, I shouldn't see it).
The Semantic Web ideally means that the data you put out on your web site becomes machine readable, meaning that computers can read your website and act on the information as well as human beings. For instance, if I wanted to call Joe's Pizza Restaurant and I know that his website is joespizza.com, I should be able to just type joespizza.com in to my phone, and it would go to the website and get the number and call it for me. But a computer can't easily determine the difference between a telephone number, a fax number and his opening times. They're all just numbers. That's why, if you explicitly mark up the information on the page, software can do things with it.
It may be that initially, you want to make your phone number machine readable so I can download it in to my phone or PDA. Then you make your menu machine readable. Then you hook up your stock control to your site, so people can find pizza restaurants with enough ingredients to make a pizza to my requirements. And you geolocate it so that I can plan my route to the restaurant based on the data provided by the local transit provider.
[edit] But, why don't I just lie?
Well, you could lie. But what does that do for your reputation. If you say on your web page that you serve spicy vegetarian pizza and you serve only spicy chicken pizza, I will be quite unhappy when I've walked across the city only to find you don't serve the type of pizza I explicitly searched out for. Lying about your products on the Web is an utterly silly thing to do, and the Semantic Web is just a machine readable version of that.
If you lie about things like that, your customers/irritated would-be customers are fairly likely to write bad reviews of your business on their blogs and tag it up specifically so that the next time someone searches for Joe's Pizza, they are going to find a rant about how Joe lies about his pizza toppings. Instead, they may have written a nice comment about your pizza if you didn't lie.
Data you put on the Semantic Web may be thought of as analagous to certain data you put out anyway. So, for Joe's Pizza, it's equivalent to an advertising flyer or downloadable menu on their site. Lying on the Semantic Web could have the same kind of real world consequences for a business as lying in real life - irritated customers, legal action or, for instance, a body like the Advertising Standards Agency breathing down one's throat about false advertising.
This may sound somewhat absurd, but it's an argument that is often brought up (see the Arguments against the Semantic Web page).
[edit] Why should I bother with this? Nobody else is.
That's a fair point. It is becoming easier and easier to add semantics to one's site, and there are plenty of people to help you do it. "A little semantics goes a long way", as some people have said. You can add very small amounts and provide enormous benefit to your users.
For instance, the hCard microformat allows people with certain types of computers to automatically add your name, address and other contact details to their computer's Address Book with ease. It would take a competent web designer all of fifteen minutes to add it to their site, and it makes it a lot easier for what is currently a small minority of people.
But as more sites start using it, it will become more useful. It is a symbiotic cycle between the producer of the content and the user.
[edit] Someone might use this stuff to hack my site or steal my content.
This is a justifiable concern, and one that non-technical people should take seriously. There are a number of ways to go about adding semantic data to one's site, and they offer different risks and rewards.
For instance, the embedded semantics route - microformats, eRDF and RDFa - do not present any real kind of security risk at all. The external semantics route - providing a separate file, or what is called a SPARQL endpoint - may provide some security concerns, but these are easily solved by experienced people, and should not be enough of a concern to cancel such a project, but rather just to make sure that the people implementing it are knowledgeable about security.
As for people infringing on your intellectual property - yes, that is another concern that people do have, but it's one that should not really be significant. If you put data out there on the Semantic Web, it is at as much risk of copyright and IP infringement as any other data that is on the public-facing Internet. There are mechanisms - the 'rel-license' microformat and the cc:Work class in RDF which allow you to provide a Creative Commons license on your data to allow other people to use it. But providing data publicly in more semantic formats does not imply that you have given up any IP rights.
