Metadata, Resources, and the Resource Description Framework 5 attributes are limited to those that make sense in the world of books, and not for a more general set of knowledge representation, or even for magazines. The  knowledge  representation  community,  on  the  other  hand,  was  the  hot thing of the late 1980 s, expected to become the next huge thing in the software industry. It fizzled for different reasons, but they came at the problem from a different direction. An expert system is a set of rules, into which you can fit knowledge as input, and get different knowledge sets as output (for instance, if the input is the symptoms of a patient, you can get a recommendation for treat- ment as output). But this means that you want to find a way of representing any and all knowledge that fits into the system. This brings a different set of problems. If you just think about metadata within one single domain, it makes sense to have a table with the attribute names in one column, and the attribute values in the other. But if you want to compare the attributes to something else, how do you know what they represent? What is the unit of  measurement if thickness is 1, for instance? Inches or centime- ters?  Or  millimeters?  Or  meters?  Or  miles?  This  seemingly  simple  problem made the Mars Surveyor crash, and it can easily make someone else misunder- stand what you want to present, especially if you want to present it to a com- puter program. A computer is stupid, and if you do not tell it exactly what you want  to  say,  you  cannot  expect  the  receiving  software  to  take  any  relevant actions. RDF, the Resource Description Framework, was developed by the World Wide Web Consortium (W3C) to create a format for making assertions that leverage the XML format to represent and transport information. XML, however, is not a markup language; it is a set of rules for creating markup languages. I will talk more about this in Chapter 2, but suffice to say that using XML brings us as near to a universal data format as is possible nowadays. However, XML gives only the rules for how the byte strings should be cobbled together to form a coherent whole, which can be used by a widely spread set of computer programs. How an XML data set should be interpreted is determined by the Document Type Description (DTD), which essentially creates a schema for an application of XML. But XML does not say anything about the informa- tion itself, only the way it is structured. The level above XML determines how the information is interpreted, and this is where  RDF  exists.  It  builds  to  a  very  large  extent  on  experiences  from  the knowledge representation community, creating a way of comparing different knowledge representations by making the user define them in a specific way and enabling the comparison to use a format which is defined in a branch of mathematics called graph theory. RDF is a way to express relations between objects, something XML does not allow you to do. 69528_CH01Ix  4/6/2001 8:15 AM  Page 5