Semantic Predictions 2009
After recommending the “Predictions for 2009” of the major experts in the IT field, now I’d like to add my own point of view on future scenarios for semantic technology. To be honest, these are expectations more than predictions: if any should come true, I will consider them as fulfilled predictions, or if nothing should happen, they will just remain as wishes:-)
1. Renewed attention to the quality of technology and concrete and measurable applications. This first wish may be an easy prediction, as recession always tends to mark the end of nonsense and Pindaric flights, i.e. flights of fancy proliferating in times of plenty (just think of labels such as Web 3.0 and all sorts of speculations on Semantic Web and semantics of all kinds). I very well know that the hype of semantic technologies won’t last forever (a new bubble is always waiting around the corner) but, as we are currently in a recession, at least for the next couple of years we won’t have to see 10-20 million dollars wasted in start-ups lacking innovative technologies and a concrete approach to the market (created only to get the attention and money of some big fish ready to buy them for fear of missing opportunities, thus feeding a vicious circle and giving sense to a business model that makes little sense in itself).
2. A more widespread and concrete use of semantics on the Web. Again thanks to the processes mentioned above, hopefully we will start to see some genuine use of semantic information: for example, using conceptual information, and not only keywords, to retrieve information in a more accurate and correct way, as well as relational information, i.e. relating data that are not directly connected. Such developments will come as soon as the experts realize that manual tagging and folksonomies (however valuable and often useful) do not represent a true solution to the problems of knowledge management on the web, because all things considered, like everything done by human people, manual tagging and folksonomies are too subjective and dependent on individual abilities and will.
3. Linked-data: a new semantic trend on the Web. I expect linked-data to become the next buzzword for those dealing with semantics on the web. Linked data are easy to develop (at least in part), contain the concept of link, and are great to imagine fantastic scenarios where everything is saved and cross-referenced. In reality, at least in their present implementation, linked data aren’t particularly innovative: the possibility to start from a text where Steve Jobs is mentioned and follow a link to his article on Wikipedia and maybe to some other data extracted from a well-known web site, doesn’t seem very innovative to me.
Something different (and truly revolutionary) would be the possibility to have these kinds of links available also for people, companies and products not universally known nor widely discussed (in a semi-structured way) by authoritative sources. To do that, however, we need a true semantic technology, infrastructures (like this one also involving us) and a lot of hard work. The concept of linked data is very interesting and deserves developments, so we hope that the old unrealistic expectations (regularly disregarded) won’t make someone throw the baby out with the bathwater.