Conclusion

The semantics of a typical business application are complex and voluminous. We can't escape that. There is a certain amount of essential semantic complexity in the systems we build. That is the bad news. The good news is that the essential semantic complexity is nowhere near as complex as what we have been doing to date, which is to work around the semantic complexity. The complete semantic description of a typical business application consists of thousands, and in some cases tens of thousands, of semantic expressions. This is a lot to comprehend, as well as to maintain and configure as things change. But it is two to three orders of magnitude less complex than the current implementation of the same system, which will typically run to millions of lines of code.

In this chapter we've described the type of information that needs to be captured to translate the semantics you've uncovered into semantics that can be converted into a system. We've described major types of tools for cataloging the semantics that you've captured, as well as the major modeling approaches currently in vogue. Although each of these can be used for expressing and communicating semantics, and you will probably use the tool or approach that you are familiar with or have already purchased, none of them is ideal. I've laid out the requirements as best I know at this point, and I hope that some enterprising software developers will come forward and build something into this outline that will make semantics far easier to express and convey.

Finally, we dealt with the issues of expressing semantics formally and the role this will play in converting our semantic models into systems that actually do something.

The remainder of this book picks this thread up and explains the key role of semantics in some key technology initiatives as we enter the twenty-first century, specifically XML, enterprise application integration, Web Services, and the Semantic Web.