AI Zone Admin Forum Add your forum

NEWS: Chatbots.org survey on 3000 US and UK consumers shows it is time for chatbot integration in customer service!read more..

Is it time to create an AI standard language?
 
Poll
Is there a need for a new AI/Chatbot language
Yes - Create a new independent language 9
Yes - But, I would prefer AIML was just enhanced 8
Yes - But, it should be based on a current language 3
No - each developer is best doing his own thing 5
No - I don’t think one language cold do it all 3
No - It is a waste of time 2
Total Votes: 30
You must be a logged-in member to vote
 
  [ # 76 ]
Merlin - Oct 2, 2013:

....

Well, Python3 has some OS-specific commands, but that makes it more powerful. However, I understand your view-point. Since my Betabots are made for Unix systems only, I saw no need to make Xaiml a cross-platform language.  Plus, being system specific makes the language more powerful.

Could you explain how Xaiml “needs flexibility”. Thanks!

Xaiml is not intended to be simple. AIML seems too simple, so I made Xaiml to be the language I felt AIML should have been.

I see your viewpoint on the divergence from XML standards. That is why I usually say Xaiml is XML-like. Plus, I always got sick of syntax errors, so I fixed that in Xaiml.

True, Xaiml is very different from other languages, rules, standards, and concepts. However, why not try to make something new and revolutionary? Why stay stuck in a ditch of standards that are restrictive? Why be stuck in a mind set that X can only be done by doing Y? Computer languages need to be more “understanding” and “forgiving”. A computer language that knows what you mean is more “intelligent” on the computer’s side.

You are the first person to think having all of these programming tags (, <ruby>, etc.) is a bad idea. I can see why you would think so. However, I always felt AIML was too restrictive only having <system> and <javascript>. What if someone does not know javascript or the system’s shell?

The tags that are like that (<french_ruler>) is mainly for experimental purposes. Notice that the Betabots are in the Alpha stage and Xaiml is in the Beta stage. Some tags are intended for experimenting with different ideas and concepts.

Thanks for pointing out the Adobe tag issue. That needs to be removed. Would you like to be listed in the documentation as a contributor?

The application tags (<opera>) are intended to be shortcuts of <sys>opera</sys>. If a system did not have Opera installed, then nothing would happen.

Does Chatscript/Rivescript have special support for emotions without needing to type more than three lines? Does Chatscript/Rivescript allow spelling errors? Aliases?

Thanks for your thoughts. This gives me more to think about. Perhaps, I should make a second Xaiml standard - MiniXaiml. Thanks for the thorough review of my documentation and Xaiml specification.

You seem like a very interesting person. Do you have a Google+ account?

 

 
  [ # 77 ]
Merlin - Apr 18, 2011:

AIML ... has not changed in a decade.

Really?

 

 
  [ # 78 ]
Thunder Walk - Oct 3, 2013:
Merlin - Apr 18, 2011:

AIML ... has not changed in a decade.

Really?

Yep.
At the time I made that post, AIML hadn’t been updated in about a decade, and the spec was just a working draft.

Artificial Intelligence Markup Language (AIML)
Version 1.0.1
A.L.I.C.E. AI Foundation Working Draft
25 October 2001 (rev 006)

Of course now you have a working draft of AIML 2.0.

 

 
  [ # 79 ]

Devyn Collier Johnson: Since my Betabots are made for Unix systems only, I saw no need to make Xaiml a cross-platform language.  Plus, being system specific makes the language more powerful.
All of that is goodness for your own project, but suggesting that “Xaiml is the best way to go” as an AI standard language (which is the topic of this thread), was misguided. It looks like you designed it more as a personal/local unix application.

I think Dave put it best:

Dave Morton - Mar 14, 2013:

Given that level of complexity, may God have mercy on the individual who is tasked with writing an interpreter to handle it, especially if the tags are in foreign languages! big surprise

My prayers also go out to any botmasters who might be tasked to write the Xaiml code for a trilingual chatbot! cheese

Could you explain how Xaiml “needs flexibility”.
I did not say Xaiml needs flexibility, I said an AI standard language needs flexibility.
<javascript> is not supported and I do not plan to do so.
The one disadvantage of my Betabots is portability. They will not run on a Windows system.

Multi-platform might be one example of the flexibility required. Support for the world’s most popular scripting language (which was also supported in AIML) might be another.

In a standard, having a “Script tag” and an attribute tag indicating the language might be the way to go.

I see your viewpoint on the divergence from XML standards. That is why I usually say Xaiml is XML-like. Plus, I always got sick of syntax errors, so I fixed that in Xaiml.
If the standard was going to be TAG/XML based then it should be fully compliant. Then you get the use of all the XML tools for free. Many of us don’t like XML and the syntax errors, but instead of adding more tags, we developed tagless languages. Chatscript/Rivescript/JAIL all fall into that camp.

Why not try to make something new and revolutionary?
You should try to push the envelope, especially for your own project. It is just not viable as a “standard”.

You are the first person to think having all of these programming tags (, <ruby>, etc.) is a bad idea.
Let me be clear. I think the ability to use external programmatic resources is good. It is just your implementation I question (especially in a standard).
Instead of 3 flavors of Python tags, and 3 flavors of Ruby tags, where are the:
JavaScript, lua, R, Matlab, Basic, LISP…
C#, C++, FORTRAN…
Tags?

Better to have a “Script” tag with an attribute defining the language.

The tags that are like that (<french_ruler>) is mainly for experimental purposes.
Where is the (<french_pastry>) tag? Or the (<french_province>) tag? wink
An AI standard language could use the concept of an ontology or sets, I provided you with a link to a number of implementations in your original Xaiml:// A New Chatbot Markup Language thread. I suggest you review these for some implementation ideas.

Thanks for pointing out the Adobe tag issue. That needs to be removed. Would you like to be listed in the documentation as a contributor?
No need. I am just here to help.

The application tags (<opera>) are intended to be shortcuts of <sys>opera</sys>.
I suggest creating a standard way for a developer/user to create the shortcuts themselves, rather than trying to build them all into the interpreter. You should also review AIML 2.0’s
“out of band” (OOB) commands.

Does Chatscript/Rivescript have special support for emotions without needing to type more than three lines? Does Chatscript/Rivescript allow spelling errors? Aliases?
Each language has pros and cons. If you are serious about developing your own AI interpreter I suggest you explore the other languages/interpreters and identify features you may think are useful. 

You seem like a very interesting person. Do you have a Google+ account?
Thanks, I don’t do much on Google+ but I have been known to show up at a hangout now and then. smile

RiveScript vs. AIML
AIML 2.0 OOB

 

 

 
  [ # 80 ]

A while back I started a Google+ community for chatbots.org members who might want to hangout or do video calls. A few of us have taken part in video calls from time to time since then. Feel free to join if you are on Google+ and chatbots.org.

https://plus.google.com/u/0/communities/117236447102577073584

 

‹ First  < 4 5 6
6 of 6
 
  login or register to react