Much of the information we process with computers is not in a form that computers can [process]. Source code for compilers is the major counterexample -- writing source is not a common endeavour except within this wiki community. Here, our PatternLanguage is helping us to learn to write better source code. ----- "... computers can...what" interpret? understand the meaning of? ''My suspicion is "... that computers can process." The second occurrence of "process" is then elided.'' Let's see: Financial transactions are often "processed with computers." How many computerized accounting systems do you know of where the accounting transactions are "not in a form that computers can process"? If computers literally couldn't process "the information we process with computers", then we'd really be in a pickle. No, we need to be more clear about what we're trying to say on this page. I think we're trying to say that the current generation of computers and software can't "understand" the "higher level meaning" of the data they process. Yes, a properly instructed computer can do the math of crediting my checking account and debiting my account for visa liability, but there may be no fundamental abstract "understanding" that I've just "paid my visa bill." Likewise, my word processor can display and print symbols and makes both spelling and grammar suggestions, but it can't really "understand" my letter to my mother. Is this what we're really trying to get at? And if so, is such computer "understanding" really necessary or desirable?