AI Example Help

[Trevor Nelson]

> 
> An interesting concept. AI monitoring systems have never really
> 
> appealed to me; I personally prefer something with simpler and clearer
> 
> rules (eg "if server load exceeds 3.0, raise an alert"), coupled with
> 
> information retrieval commands that read like commands, not natural
> 
> English. But I can see the value of the more "human-friendly style",
> 
> if that's the term for it.
The other goal on this project is to make the application work it's best to automatically narrow down the cause and correct it by itself with a report.



> 
> > I greatly appreciate any help you guys can prodive me with to help me get started. Before you assume I want everyone to do all the work for me, it's not the case. It's just that I learn best/faster from example code.
> 
> > Also, I am thinking of going with python 2.7 for this since it seems all the documentation I run into isn't for the 3.x Python.
> 
> >
> 
> 
> 
> Understandable.
> 
> 
> 
> I would recommend treating this as two completely separate projects:
> 
> 1) Natural language request/response - the "AI" bit, tinkering with
> 
> the HOWIE bot
> 
> 2) Server status querying, log analysis, etc, etc.
> 
> 
> 
> You may already be familiar with one or the other, I don't know. But
> 
> in any case, they're quite separate; it'll be easier to develop one
> 
> and then the other. Divide your problem into pieces, then divide the
> 
> pieces further, and keep on dividing until you have a series of
> 
> one-banana problems[1].

This is what I have actually been thinking of doing is splitting up the HOWIE bot.  But, that is one thing I am working on still to really figure out how to split it apart between automagicness of telling you things without user input.  And then the input and answer part.  Although, from what I can think of at this time is to write a main program that is the "Automagic part" and then it can call in Howie as needed?
> 
> 
> 
> I would also recommend using Python 3.3, for several reasons:
> 
> * There's plenty of docs for 3.3; perhaps web searches are
> 
> highlighting 2.x, but if you go to the main docs page[2] you'll find
> 
> 3.3 there.
> 
> * Py3 has much better Unicode support than Py2, a significant factor
> 
> when you're working with natural language parsing
> 
> * The 2.x line is no longer being developed. Only bugfixes and
> 
> security patches will be applied. 3.x, on the other hand, has cool
> 
> toys added periodically with new versions.

I will have to see how HOWIE runs with python 3, and give it a shot.