Hey there Neil, I got this main and kind of hooked up what you were trying to tell to this Alan Grimes. What you did was explained what we've currently got in praxis. This might be good in convincing somebody like Harold we have done something, but for as up-in-the-clouds AI devoloper as Alan Grimes, it simply won't proove a thing. As he nearly said it himself: You need to impress him with something. Remember Lisp? Well it's just another programing language, but because somebody says it has something to do with AI, everybody that knows about AI also knows about Lisp. Here is the plan: You too need to lift the discussion of the ground a bit, because if the abbilities of our system were so obvious from plain evidence, we would have AI at home for about 20 years already. Tell him that we've got a network platform intended to make the Internet equaly friendly to users and software. Tell him that our network is the perfect environment for any type of aritificial mind. Tell him that we have included techniques in the very mainframe of our networks that takes software away from the hardwired world and into the endlessly flexible world of AI; think of the ICI script language, how it can be automaticaly built on the run without any complications of ordinary programing languages. And don't forget to tell him that our network platform is built in a way that it dosen't need to be finnished to function, meaning that we should have some level of AI very soon. He needs to know that this network platform offers a programing language, a network, an unlimited number of functions, ultimate compatibilty and ways to link all of this in you-name-it ways, providing ways to hybridize existant 'AI' engines all making one supreme global and immediately usefull AI: the Core. This should get his wheels spining and he will eventualy get our ideas and our progress. Because as long as he thinks the way he writes, he's as good devoloping AI as he is discussing it on . Alan Grimes wrote: > In VB!!! =P > > Look: Writing stupid calculator programs does not advance the field of > artificial intelligence/cybernetics. Unless you are working on something > significantly more advanced, I won't be interested. Neil Nelson wrote: > You will see one of our software author's (Jure Sah) related Internet > communications project at http://aimetasearch.com/ici/ICICH.html. > That project is in VB and the web page is new as of two days ago. > > I am writing the secure Internet communications layer ICSA in C/C+ > that will run on both Windows and Linux. This layer currently has a > very simple file interface method that will provide a TCP > communication ability for any system in any programming language. > Faster interface methods are under consideration, but unless Internet > communication speeds can increase quickly overall, the delay of a > file interface method in the short run (files remain in memory if > accessed quickly enough) should be a minor consideration in the total > communication delay. > > Harold is writing system administration software in a third language > and has a strong interest in the supercomputer or Internet parallel > computing (IPC) area. I also have some projects--general NLU being > a primary one--I would like to run on the Internet parallel computer. > > It seems to me that a required key area for the Internet parallel > computer, unless Harold has this area well in hand, is some specific > method that can be used by programs to enable distribution, parallel > execution, and return of the partial results (and with iterations of that > basic process). On a Beowulf Cluster this area is covered by MPICH > that consists of a set of routines and calls in C. MPICH appears to > be a layer on the socket method geared toward the supercomputer > parallel method and communication requirements. > > And so there needs to be a well organized specification for this IPC > enabling area and then some software development coordinated with > the ICSA and Harold's software. One way to go about this is to > identify a simple example of software that might be run on an IPC > that contains most of the elements any IPC software may require and > then work up a design that will allow that simple example to work, > and then integrate that design with the other components, and then > code the design. > > There are a number of projects being considered, and we are very > interested in working with software authors in their area of interest. > If you have some ideas along this line, please let us know. C'ya! -- Cellphone: 0038640809676 Don't feel bad about asking/telling me anything, I will always gladly reply. Digging for info? Try AI Meta Search: Http://WWW.AIMetaSearch.Com MesonAI -- If nobody else wants to do it, why shouldn't we?(TM) Http://WWW.MesonAI.Com