Principles Imply Patterns

In biological multicellular systems, the DNA that directs each and every cell encodes mechanisms that support each of the principles. But that happy circumstance is a result of at least half a billion years of rigorous evolution. Organisms that didn't get it "right enough" didn't survive. Since we don't want to wait so long to get it "right enough" in multicellular computing, we need to let "intelligent designs" get us there quicker. For that, we use patterns to tranmit design principles.

But multicellular systems don't arise in just one circumstance or at one scale. Circumstances and scale will determine the degree to which design dominates evolution, or vice versa. For example, relatively small systems that run in one, or a small handful of machines are primarily designed. Large IT systems that involve hundreds or thousands of machines make a good pretense of design, but tend to be overtaken by evolution before they ever go online. And systems in the Internet, e.g., P2P systems, clearly are dominated by evolution.

Think of the FourPrinciples as something like Asimov's three laws of robotics. Necessary for multicellular systems to be housebroken, but hardly sufficient. And, just as Asimov's laws played out in subtlely different ways in each of many science-fiction novels, the four principles will play out differently in different systems.

Therefore, patterns for small designed systems will very likely not translate well to Internet-scale evolving systems.

Since the IntertwinedPrinciples need to work well together as well as illuminate the issues separately, the architectural patterns will likely emerge in sets of five: one for each principle and one about how their functions are intertwined.

Until good patterns emerge and become documented, consider as an example, TheRuleOfHand

Last edited June 29, 2006
Return to WelcomeVisitors