Hi, we are exploring building our own in-house AIML server at some point.

I was wondering if people have had experience or could direct me to a resource to help us to start to think through the design and the logic.

Would be curious around other people’s experience.  Thanks!

What, exactly, do you mean by an “AIML server”, Tim? That phrase could mean several things, really. For example, it could mean software that supplies AIML categories or files to an AIML interpreter (such as Program AB, or Program O) from a remote location, or it could mean the AIML interpreter/chatbot engine itself. If you mean the latter, I can tell you from personal experience that it’s not the most straight-forward of tasks. I’ve been working on Program O for many years now, and it’s still far from what I envision for the project. don’t get me wrong, I love working on Program O, but right now it’s just me working on the project, and between time constraints, and the lack of people to discuss techniques, methods and direction, the project has been foundering a bit. It doesn’t help that, since my bike wreck back in 2012, I can’t seem to focus on things for more than a few minutes at a time, but such is life.

What are the differences between a program that works on the desktop for one person and a program that works in the cloud for thousands simultaneously

- Multitenancy
- Multitier architecture
- Multithreading (computer architecture)

If an AIML interpreter works on the desktop for one person, would not a cloud based AIML interpreter that works for many simultaneously be a “server”?  Dave, how is this handled in Program-O?

- Technically, how does Cleverbot manage talking to 100,000 people simultaneously
- How many people can Siri talk to simultaneously, and how does it manage this technically

Marcus Endicott - Nov 26, 2014:

If an AIML interpreter works on the desktop for one person, would not a cloud based AIML interpreter that works for many simultaneously be a “server”?  Dave, how is this handled in Program-O?

Well, Marcus, Program O, while being “cloud-based”, is really platform agnostic. Programmers of desktop apps can easily set up a web interface to interact with a Program O chatbot that’s remotely hosted. I know, because I whipped up similar experimental interfaces, not only for Program O and Pandorabots, but also for ChatScript, and I’m a very “entry level” Visual Basic programmer.

I think that the term “AIML server” is a bit of a misnomer, given the context of the original post, but until I get some clarification I’m not completely sure of that. I guess we’ll just have to wait for Tim to come back with a bit more insight into his needs.

Thanks, Dave.

I am looking to build something that would parse AIML, so, yes, the interpreter for our own internal use.