Why URIs for statistical classifications
In all statistical classifications we can distinguish a terminology and a coding system. Statisticians are usually mostly interested in the coding part, as codes are what is used in tables and databases to serve as dimensions of the data. Codes are more practical then names, as they are (usually much) shorter and easier to manipulate in a database. However, codes alone do not ensure that data is unambiguously described, as the "meaning" of a code is only defined within the scope of the classification it belongs to - once extracted from its classification of origin, a code is ambiguous, at best, if not meaningless altogether. In this sense, codes are "local" entities, and raise interpretation issues when passing data from one system to another. Moreover, also the terminological part plays a crucial role in statistical classifications.
Caliper aims at making the sharing of statistical classifications as smooth, unambiguous and interoperable as possible. To this end, it provides classifications and items with unique identifiers over the web, aka URI (Unique Resources Identifiers). A local code is then paired with an URI that allows one to unambiguously identify it - always and everywhere. URIs are a especially useful for machines (computer applications), as they can access the corresponding web address (URL) and retrieve all relevant information associated to the item (in a language that machines can process automatically). Humans may also access those web addresses and see a simple page showing the same information accessed by a machine, but more nicely visualized, suitable for human reading. This mechanism of serving data differently depending on who access the web address, is called "content negotiation", a standard and well established web technique. Caliper implements content negotiations through Loddy, an open source software powered by the University of Tor Vergata.
For example, the URI* for "wheat" in WCA is "http://stats-class.fao.uniroma2.it/WCA/crops/Wheat". That URL will resolve in different ways depending on "who" access it. To a machine, it will appear as just a bunch of RDF statement, while if you use it your browser (say Firefox, Chrome or IE), you will see this page:
The tool we use to render the RDF into an HTML page is Loddy.
Try the following URLs:
- http://stats-class.fao.uniroma2.it/WCA/crops/Wheat is the URI for "Wheat" in the list of Crops of the World Census of Agriculture;
- http://stats-class.fao.uniroma2.it/CPC/v2.1/0118 is the URI for "Millet" in CPC 2.1.
URIs for concept schemes
Some users may find useful to have the full list of URIs concept schemes currently included in Caliper (run this query to get always the current ones)
(*) NOTE that, being Caliper an experimental service, not yet an official service of FAO, these URI are subject to change. They should not yet be used as reference.