That thought crossed my mind a couple of times!! Yes, I think a thread per bot   No problem about being strict. . ..otherwise, the threads will be a mess !

Hi,
I’m not personally bothered by all of the posts in this thread. However, it does make sense to have separate threads for various bot projects.

*Update*
Envious of Vic’s PERL pattern matching capability I started playing with ‘regular expressions’. VBScript supports this…so I can test in VBA - Excel.  I also located a module that may be used in C++. I hope to give that a try…that’ll certainly make things easier with Walter.

Regards,
Chuck

Hi chuck… judging from your personality I figured you wouldn’t be. . . I may have hijacked (or so it would seem) the thread from time to time, but I think you have gained some new ideas from reading the posts

Vic,
Oh yeah!  I’ve learned a lot here and appreciate the constant feedback and various perspectives.

At the game dev site we all start our own threads to document our game projects during each competition.  This works very well.

Regards,
Chuck

Chuck,

I’ve really enjoyed following your progress on this thread (and the side trips into everyone else’s work as well!).

Your approach made me think of a project from the late 60’s called SHRDLU (sort of the Eliza for this approach). Basically, it was a bot which would interact with a simple world composed of blocks with different sizes, shapes, and colors. It would take commands to move blocks around and answer basic questions about its environment.

Hi Justin,
Thanks!  I hadn’t heard of that 60’s program you’re describing. I would certainly be interested in learning more. I’ll Google that later.

Aren’t side trips awesome?  There are always a few nuggets along the way. =)

Regards,
Chuck

Building a chat bot is difficult to say the least!  =)

I’m in the design phase of this little project and I have two supporting test programs.  However, the number of things I need to consider seem to increase on a daily basis…such that it’s beginning to feel a bit overwhelming. I’m certain reading about AI is a good thing, but it does send my design considerations off tangent.

I’m going to revisit my program spec and compile a “to do” list to help real this in.  One thing is for certain, I want to get basic chat with Walter going soon…even if it’s very simple and limited. Then I can post the program for others to download and test.  That feedback will be most helpful…and encouraging.

Regards,
Chuck

I’m sure I speak for the others when I say that we’re behind you, chuck. I, for one, would be tickled pink to help test out the program, once you feel it’s ready for us to do some meaningful investigating. Just let us know what you need, and I’m sure that between us, we’ll pitch in and, “Get ‘er Done!”

By the way, how’s the game contest going?

I can also help out testing. . .even if it means installing windows on a old box in the basement, or in a virtual machine on top of my Ubuntu Linux:

Thanks guys for the support!  Vic…I’m honored that you would consider installing Windows just to test my itty bitty program.

Top of my ‘to do’ list is defining a file structure for question and response patterns.  I spent a couple of hours thinking this through and I’ve come up with a format for the files and a method of processing within the chat bot.  I’ll post more when I’ve got more to share.

Dave, regarding the game, so far so good.  I populate a window with various 32x32 pixel tiles. Each tile has part of a song attached to it. The goal is for the player to sort out the musical bits and to arrange them in correct order.  I’ll post a demo for you to test if you’re interested. The cool part is the sound is not heard until the mouse moves near the center of the tile. Sound fades when approaching or moving away from the center.

Regards,
Chuck

Hi Chuck, begin away for a while and the forum, make it a little hard so go through all postings. Considering this, it’s probably also hard for new visitors to understand what’s going on with Walter. Last time, I was involved in all geo-stuff, now it’s totally different again. I’ll do the best I can to get back on track . I’ll back, promise !

Chuck, it would be my pleasure to help out.  It turns out I do have a machine here, with XP on it and .net runtime 3.5.  Does Walter/Carolina run on .net runtime or ‘bare metal’ cpu ?

Erwin,

Last time, I was involved in all geo-stuff, now it’s totally different again.

A programmer with Attention Deficit Disorder….that’s all I can say. =)

Actually you’ll be pleased to know that I’m working on adding chat about Walter’s world. I simply needed time to sort out the initial design issues.

Vic,
I’m using Visual Studio 2005 to compile my C++ code. I’m not using the .net framework.  ‘bare metal’ CPU? Never heard it said that way. Sounds cool!

*Update*
I needed the ability to embed global variables into output responses.  Downloaded boost.org libraries…specifically to use the ‘regex’ (regular expressions) code.  Here’s a sample demonstrating a predefined output response and variable substitution.

So, the human questions will be translated into a standard expression and unique function. This function will retrieve the global variables and substitute the values in the output.  This boost code will make things much more simpler.

Now I need to create two files…input pattern questions and output responses.

Regards,
Chuck

I’m not sure where I picked up that expression….. running on the bare metal (meaning compiling right to machine code, rather than compling to MSIL or java byte code to run on top of a virtual machine or run-time).

Anyway, I tried the whole .net thing and I found it was just another pain, similar to compiling java.  So called “write once, run anywhere”... yeah right!!!  I really never got the logic… If you compile your C program to machine code for say intel, then it won’t run on a SPARC, true, so let’s make java so you compile to ‘byte code’, then you can run the byte code on any cpu. . .but you still need to re-compile anyway!!??  Maybe just me, but i don’t get that

  It is really nice to know there are people like you out there that still code in C++ and compile to machine code and haven’t been ‘assimilated by the .net/java borg’ (another of my geeky star trek terms lol).

I’m using Visual Studio 2005

I’m not certain from which version Microsoft implemented this (2003, 2005 or 2008), but, if you don’t do some serious drastic project reconfiguration in C++, it will also produce the same .net il code, but (I believe) with the extra possibility to compile further to native binaries (which is done automatically, I think). There is no difference any more between the code rendered from a C++ and a C# app (or vb.net for that matter) because of the simple reason that it is the same compiler.
@Victor: you are absolutely correct that a .net application needs to be recompiled for a different CPU architecture. It’s even more extreme: if you install a .net (C# or VB.nt) app in a regular manner, you only install the compiled IL files, not real exe’s. So the app needs to be recompiled each time that you run it (that’s the jit: just in time compilation). You can install in a different manner, so that it compiles to your machine architecture (paint.net, which is free, does this).
The main advantages of this technique are:
-the IL files are generally a lot smaller compared to exe’s because intel instructions require a lot more space compared to IL’s.
-The unified compiler is handy for microsoft to maintain: only 1 compiler instead of 4 or 5.
-there are probably others I can’t remember….