From: rw@shadow.org.uk   
      
   "Willem de Ru"  writes:   
      
   >>   
   >> What do people use these days? (And "Matlab" is *not* an answer)   
   >>   
   >   
   > Thanks for this very informative thread.   
   >   
   > In terms of Fuzzy Logic code libraries: I was just wondering why Matlab   
   > is not an answer? (I am not asking because I necessarily disagree, I   
   > just want to learn from other people's experiences as well.)   
      
   I'm doing embedded development, prototyping, real time, Free software,   
   and running under Linux. As far as I understand, Matlab doesn't have   
   anything to offer me :->   
      
      
      
   > With Matlab I have the following: It provides me with fast matrix   
   > programming (which can e.g. very effectively implement a FAM) to speed   
   > up my development effort; Matlab serves as a central design interface   
   > where I can utilise e.g. (1) "Pure" Statistics functions (for e.g.   
   > pre-processing), (2) Neural networks and (3) Fuzzy Logic systems,   
   > depending on the type real-world application to solve. If I had to do   
   > all this in C++ (or similar language) it would probably take me much   
   > longer to develop? If required, I can use the Matlab Fuzzy Logic,   
   > Neural Network and Stats Toolboxes to "fast-track" the development of   
   > my systems, but I can develop it from the ground up as well.   
      
   Yes, I understand that it works for people. However, I tend to find that   
   graphical interfaces get in between me and the system...   
      
      
   > The benefits of C (or similar) libraries/code are that I have lower   
   > level control over exactly what happens in my system and I also   
   > probably have processing speed advantages. Moreover, it is easier to   
   > incorporate it into eg. systems that also have been developed in C (or   
   > related language). And in many situations integration is very   
   > important. In C (or related language) I have more flexibility in terms   
   > of developing a user interface, database access, etc.   
      
   The integration angle is also pretty important. It's hard to integrate   
   something you don't have the source code for.   
      
   cheers, Rich.   
      
   >   
   > Just wondering...   
   >   
   > Regards   
   >   
   > Willem de Ru   
   > BTW Earl, thank you very much for the "Fuzzy Modeling and Genetic   
   > Algorithms for Data Mining and Exploration" book! It is an excellent   
   > addition to my library! I learn a lot from it. Thanks! Your books have   
   > been *very* important to me for helping me to keep the faith and   
   > excitement in Fuzzy Logic.   
   >   
      
   --   
   rich walker         |  Shadow Robot Company | rw@shadow.org.uk   
   technical director     251 Liverpool Road   |   
   need a Hand?           London  N1 1LX       | +UK 20 7700 2487   
   www.shadow.org.uk/products/newhand.shtml   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)