Developing a standard online wine sommelier ontology
Now that wine is making its way into the online community, folks have begun to tinker with the idea of an online wine sommelier. In theory, it sounds like a wonderful idea.
You grab your iPhone or Blackberry, connect to a site, and enter your dinner choice. The mythical program then determines your location via the GPS function of your handheld device and hunts up the online version of the wine list or wine inventory at your current location. The program then cross applies your specific taste profile and your meal information and recommends the best wine.
What a great idea! However, there is nothing even close to this available yet on the internet. The programs we’ve experimented with fall into two categories. Programs that require the end user to have lots of knowledge about wine to use, or programs that return information that is of little value unless you know a lot about wine.
The need for an ontological map of wine
Programmers that are trying to construct our mythical sommelier application face several challenges. First, there is no generally accepted ontological mapping for wine, using the idea of the OWL Web Ontology Language.
The OWL Web Ontology Language is intended to provide a language that can be used to describe the classes and relations between them that are inherent in Web documents and applications.
Currently, wine searching relies almost strictly on keyword searches, weighted by the numbers of incoming/outgoing links, which leads to a problem of scale.
The World Wide Web as it is currently constituted resembles a poorly mapped geography… The sheer mass of this data is unmanageable without powerful tool support. In order to map this terrain more precisely, computational agents require machine-readable descriptions of the content and capabilities of Web accessible resources. These descriptions must be in addition to the human-readable versions of that information.
Even though the idea of an OWL Web Ontology Language has been around for some time, and wine is a standard example used to demonstrate the use of ontology in programming, no widely accepted (and employed) wine ontology has emerged.
No ontology leads to inflexible searches and results
The absence of a robust ontology for wine, as well as the reliance on results driven by keywords and link popularity, lead directly to a second problem. Sites or applications that present the results of aggregated wine information do not have a flexible way to search through or present the information to end users.
For example, CellarTracker.com has over 700,000 wine reviews. But to access the data, you must enter the name of a specific wine. This is fine if you are just comparing tasting notes. However, if you want to select a wine for dinner, the vast collection of reviews has little value. To pair food and wine, you must enter wine names individually and manually sort through the reviews, hoping the reviewer has recorded food pairings in their review.
Another aggregating site is Snooth.com. Programmatically, Snooth goes through online wine reviews from end users, wineries, winemakers, and retailers to produce as complete a listing as possible of information pertaining to the wine you have entered. Snooth will also use Wine.com and Wine-Searcher.com to locate the wine.
Unfortunately, this service also requires the user to know the name of a wine to use it. If the user wants a wine that meets certain criteria, he or she is limited to price and location. To use our food and wine pairing example again, you cannot ask for a wine that goes with barbecued ribs.
Perhaps Snooth will or is planning to move in this direction. But collecting data from previously written reviews can only get you so far. They are still limited to the basic keyword/link idea, and there will likely not be enough information for a search accurately to return food and wine pairing suggestions.
A programming call to action
As members of the online wine community, we think that this is a problem that can be solved, if we are willing to work together in the spirit of the open-source model. So WineEnabler.com is looking for interested wine bloggers to work together with on this problem. If you are interested in joining us in discussing the development of a standard, open-source wine ontology, let’s talk in the comments or email us at:
Check out these related posts:
{ 1 comment… read it below or add one }
What a superb idea!
Leave a Comment