Mike, On 9/19/08, Mike Tintner <[EMAIL PROTECTED]> wrote: > > Steve:question: Why bother writing a book, when a program is a comparable > effort that is worth MUCH more? > > Well,because when you do just state basic principles - as you > constructively started to do - I think you'll find that people can't even > agree about those >
No agreement would seem to be needed - just list all of the approaches that sometimes work. Consider: ... 8. Repair is usually done breadth-first, evaluating various approaches before expending lots of effort on any one. 9. The best (and worst) methods are that way because of the domain, problem, and repairman involved. Hence, there will never be agreement among repairmen across domains. 10. The value of any particular principle is in how it works. Go with whatever works. People and computers should (mentally, physically, or computationally) try various approaches until a good one emerges. - any more than they can agree about say, the principles of "self-help". > OK, let's look a self-help as a repair domain. If you look across many systems of self help, you will see some basic truths: 1. While they have many different names, that there are only a small number of distinct types, e.g. Buddhism and Scientology have a LOT in common. There are many versions of "12 Step", etc. 2. They are easily separable along major lines, e.g. those that are for people with an internal locus of control (e.g. Buddhism), and those for people with an external locus of control (e.g. Christianity). Buddhism will never work for people with an external locus of control, and Christianity will never work for people with an internal locus of control. 3. Each of the methods, while complex in the whole, consist of small and simple steps to take as situations (combinations of symptoms, sub-conditions) dictate. For example, the first step in most 12 Step methods is to recognize a power greater than yourself. Many people are unable to get past this first step - but then again, they usually are not good candidates for 12 Step for other reasons. Hence, while we can't blindly agree which is best, I (or a computer) could ask you a few questions like "Do you believe that you control your life, or do you believe that your mother, the government and/or God is in control?" to determine locus of control, and select the most appropriate system. If they can - if you can state some general systems principles that gain > acceptance - then you have the basis for your program, > How does acceptance have anything to do with a basis for a program? It's all in the knowledge and NOT in the programming, so different principles only mean different different knowledge. The only (easily identifiable) basic assumptions that Dr. Eliza relies on seem to be: 1. That statements of symptoms can usually (perfection is NOT needed) be recognized by advanced (variable with negation and timing recognition) shallow parsing methods. 2. That applicable problems to solve have cause and effect chains (to traverse and interrupt). 3. That no traditional "computation" is needed, other than sometimes invoking "canned" programs to compute things. 4. That people will actually bother to create clear problem statements. 5. That people will actually use the program. > and it'll cost you a helluva lot less effort. There is a year or so of effort either way. One way I get a book to try and sell, and the other way I take on Google. Both ways seem to have their obvious impediments (e.g. will anyone buy such a book) and require comparable efforts. However, with a program, there is more fun and a possibility of a really BIG win. Thanks for wringing my thoughts out. Can you twist a little tighter?! Steve Richfield ------------------------------------------- agi Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/ Modify Your Subscription: https://www.listbox.com/member/?member_id=8660244&id_secret=114414975-3c8e69 Powered by Listbox: http://www.listbox.com
