From: earlcox@earlcoxreports.com   
      
   "Bill Silvert"  wrote in message   
   news:1118483215.477740.100330@g44g2000cwa.googlegroups.com...   
   > I think Walter put his finger on the issue. Earl asked why no one is   
   > asking questions like,   
   > "I am trying to represent portfolio concentration levels according to   
   > investor suitability and safety constraints but how do I use elastic   
   > fuzzy overlaps to..." and the answer is that most of us would not be   
   > interested in such a specific question. I am sure that if I posted   
   > questions about topics that interest me, such as how to use fuzzy   
   > categories to describe sediment profiles under a fish farm, there would   
   > be very little (more probably no) response.   
   >   
   > It seems to me that a group like this is best for discussing general   
   > issues which most of us can relate to, and of necessity these will be   
   > pretty academic and abstract.   
   >   
   > I also think that fuzzy logic has matured to the point where questions   
   > about portfolio concentration levels are more likely to generate useful   
   > feedback in groups dealing with finance than here.   
   >   
      
   Bill,   
      
     You raise an interesting point, however, I am not necessarily advocating   
   questions about how to design a portfolio optimization model using fuzzy   
   logic (well, not completely!), but rather discussing the use of fuzzy logic   
   in a variety of real-world contexts. This is a news group about fuzzy logic,   
   but 85% is spam and the remaining 15% is the same old discussions over and   
   over again. I just think that newbies, students, and those who are looking   
   to fuzzy logic as a solution to difficult problems would benefit from a   
   forum where practitioners bring their experience in implementing real-world   
   problems in illuminating the nature of fuzzy logic and fuzzy reasoning.   
      
      Personally I think discussions about the use of fuzzy models aimed at   
   sediment profiles under a fish farm would be VERY interesting and open the   
   door to discussion about how linguistic variables are formed, how degrees of   
   overlap affect the model, how rules are used to define the profile terrain,   
   etc. The fact that application level models could now days be discussed in   
   application forums should not exclude their discussion in this forum. It's   
   not a Boolean partition.    
      
      Just a thought.   
      e a r l   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)