John Rose wrote: > >for this job. I'm not going to use windows, for obvious reasons.
> I've seen this before, the ruling out of Windows for this type of task. > I disagree, Windows IS the OS to use for many reasons. Although some > may rule out the use of Windows for philosphical ideology. Despite the > bad publicity towards MS, MS has been very open about their > technologies. The argument used to be "Use Unix because Windows can't > handle it." Windows CAN handle intense realtime applications and with > the slew of other advantages brought along with it the decision should > be an easy one. 1. overhead. (windows is a very heavy OS). 2. configuration managment. ( Microsoft changes the specs often for the purpose of pissing off third parties). 3. Risk managment. ( With the vast complexity of windows, it is very difficult to control what your AI does and does not have access to. On a small auditable system, EROS perhaps, it is possible for a small team of human developers to excersize some measure of control over what the AI can and can't do, for a while at least. > This virtual desktop problem should actually be a minor hurdle in > relation to your other objectives. If using Windows this > can be approached in different ways. One way would be to use the > inherent desktop objects built into Windows and interact with them. That could work later in the AI's development, (and for the crudest neural interface I would sign up for), however for the time being I need to ensure that the AI generates an ontology that is intelligable enough so that the human teacher can properly educate the AI. To do this, I beleive some very minimal but still equivalent of human visual perception is required along with the equivalent of the keyboard-and-pointer input mechanism. Ofcourse we can do away with simulating hands and keys, but rather wire the software switch representing the key directly into the AI's output "cortex" (the logical equivalent of a cortex}. I have begun to look into implementing this and the problems seem quite significant from my current level of programming experience. > Another method could be to graphically snapshot the desktop image in > response to desktop changes. The analysis of these snapshots would be > the intimidating task. I have been approaching the same paradigm, You wouldn't want the AI to have the WHOLE desktop but rather an instance of the user interface in a highly sandboxed and controlled environment window. If I didn't care what the AI did with the interface, just using the raw screen would be an option. In practicality, the user must retain seperate and superior control over the underlying system. > >Software is too hard to develop. The curent state of software is that > >there are high barriers to entering the circle of developers. > >Furthermore once one has made a comittment to devel software you are > >faced with C++ which I have been told requires 7 years of daily > >experience to truly master or with the nightmare of getting a better > >language to work with an operating system which is literally built to > >support C and C++. Switching to a better OS makes things worse because > >once you get out of the mainstream you can't find support and you find > >yourself having to write/port more of your applications. While the OS > >you choose may be much much better than linux (which isn't hard at >> all)such as BeOS your costs will go up because you have to do alot >> more > Software is less difficult to develop if you leverage the existing work > done by previous developers. That is the conventional wisdom, yes. However I feel that the complexity of the existing system is so extreme that it would be better to start from scratch with a fresh design. It will be expensive and that is what the bulk of my proposal was to address... My strategy was to leverage human emotional motovators to stimulate a much more vigorous grass-roots programming effort than what can be achieved with promoting a system without an emotional hook. The plan must be bootstrapped though and that's where the $5M came in. =) > For example: do we need to write video card drivers any more? Comm > drivers? Database objects? GUI objects? Basic Math libraries? > Network protocols? Electomechanical interface boards? To get away from the horendous complexity of existing systems, the answer is a very unfortunate yes... > Do you need to do C++? No - Delphi, Kylix, Visual Basic, C#, Java, > etc... maybe try something easier. I've written some advanced > telephony switching software with 16 bit VB 3. Everyone said at the > time VB was just for prototyping and couldn't handle 8 T1's of voice. > Now VB is in Version 7. Don't rule these technologies out. C++ does > it all and it is the fastest but you may have to reinvent many wheels. > Take a look look at C#. Personally I'm doing a lot of Delphi now. > Though C++ really works well with equations - so you can do mixed > language programming and write your math routines in C++ dll's and COM > objects that are called from something easier. Or you can even hook > into the Maple or Matlab engines via COM for computability. Again, all of those have too much overhead of one form or another. Be it the price of a windows license, the tyranical EULA, the complexity of running the operating system and development environment, the complexity of the language and interfacing with the system libraries. I need a much better solution. > >Bottom line: I need $5,000,000 to start, and another $10,000,000 to go > >on-line. > If you can get $5 to 15 mil that's all fine. Many tech start-ups came > up with much more than that and flopped. With hard work and > persistance you could come up with a protoype or a toned down product > or service the generates immediate revenue for a more incremental > proof-of-concept. > These are just some ideas.... judging from your email it looked as if > you needed some... Sure, I thew togeather a non-functional interface prototype with squeak in a single evening. Putting in the functionality and getting it on-line is going to cost a lot more. The figure I gave is what I need to bootstrap this idea. (I will work out the figures in the coming weeks.) -- pain (n): see Linux. http://users.rcn.com/alangrimes/ ------- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/?[EMAIL PROTECTED]
