Versions Compared


  • This line was added.
  • This line was removed.
  • Formatting was changed.


DirectToWeb describes a technology that STILL, some nine years after its introduction, points to a logical future for how data-driven web applications ought to be developed. DirectToWeb is an expert system applied to web applications. "The general architecture of an expert system involves two principal components: a problem dependent set of data declarations called the knowledge base or rule base, and a problem independent (although highly data structure dependent) program which is called the inference engine." The WebObjects developer that wishes to use DirectToWeb to create his application manipulates the rules in the rule base to result in the desired behaviour. The DirectToWeb frameworks contain the inference engine that is used to determine what rules will be fired.

D2W AS Expert System

General information about Expert Systems below has been taken from

An inference rule is a statement that has two parts, an if-clause and a then-clause. In D2W this is referred to as left hand side (LHS) and right hand side (RHS). The LHS is used to determine whether something is true and the RHS assigns a value if the LHS condition is true. When more than one rule is true, the more specific rule (i.e. the one with more terms on the LHS) will fire. In the case of the same specificity, rule priorities are compared.

One advantage of inference rules over traditional programming is that the rules use logic and reasoning that is similar to how humans reason. The disadvantage in the case of D2W is that traditional programmers find it unfamiliar and perhaps a little threatening to think like this!

When a conclusion is drawn from the rule system and a rule is executed, it is possible to retrace the reasoning that the system used to fire that rule (there is a special debug flag in Wonder that allows you to explicitly follow the rule system step by step as it works through the logic and makes choices about the application's behaviour).

The engine that drives and evaluates all these rules in D2W is called the Rule System. Given that you'll be writing the rules that are going to be evaluated by the rule engine it may be helpful for you to know what the system is doing behind the scenes. In general, the Rule System is an inference engine with three states. The first state "matches rules" that meet the current conditions of the application. The rules that match are passed to the second state called "select rules" which determines which rules will be executed (or "fired" in D2W terminology). D2W uses heuristics to determine which rules fire. In some cases the stock D2W rule system has limitations (for example it had limited significant keys used to evaluate rules, so the programmer needed to add additional keys so that rules would fire correctly) that Wonder (ERD2W) has addressed using a more modern "select rules" process that removes the original limitations. The selected rules are passed to the third state "execute rules". The rules that execute will change the context of the application, whether through updates in the database or presenting a different UI to the user. This will cause the rule system to cycle back to the first state, ready to match the appropriate rules for the new context of the application. For much greater detail about the mechanics of the D2W rule system please see The D2W Rule System.

As an expert system that has been designed to capture the process of developing a data driven web application, DirectToWeb inherits the advantages and disadvantages of any expert system: