document.112

Electronic Commerce, Agents, and XML
We have described a number of approaches that provide automatic support in electronic
commerce ([O
́
Leary, 1997b]). Currently, the mixture of natural language, gifs, and
layout information of HTML is the major barrier for the 
automatization
of electronic
commerce, because the semantics of the information is only understood by human users.
Therefore, no real automatic processing of this information can be provided. This
significantly hampers the realization of the advantages of electronic commerce. The
information service provided by shopping agents is limited: they heuristically extract
some information, but they cannot fully understand natural language and the effort for
developing and maintaining shopping agents is costly.
The new standard XML will significantly improve the situation. HTML is a layout
language for presenting textual documents whereas XML is a language for defining the
structure and semantics of information. Therefore, it enables directed information
search and the exchange of structured data (for example, between databases). In
consequence, the automated processing of information will be possible and electronic
commerce can be executed by software agents. Still XML only provides a standardized
syntax for exchanging data. Defining the structure and semantics (i.e., the vocabulary
and its meaning), is required additionally. This is precisely what can be achieved with
ontologies (cf. [Glushko et al., 1999], [Maes et al., 1999]). We will show some of these
interesting perspectives in the context of B2B.