Hi there!
I recently discovered chatbots.org and very pleased to find an enthusiastic and friendly community and I look forward to getting involved in the discussions here.

I’ve been developing tallkigy.com for a couple of years. The intention is to provide a means for non-technical authors to publish content in chat format. I could explain why I think that is a good idea, or just invite you over to talkigy.com for a chat about the topic.

http://www.talkigy.com/chat/about-talkigy/why-is-chat-a-good-format-for-publishing-content

I love to get some feedback on how well (or not) you think this works, and any suggestions for improvements.

By the way, I have added my chatbot to the chatbots.org listing, but they seem to be stuck in pending status - anyone know what needs to happen?

Hi and welcome, David. I haven’t yet had the chance to check out your link, but it’s open in a new tab, and I hope to give it a go before dinner.

Generally speaking, Erwin is the one who approves the applications for the chatbot list, and he gets a little “tied up” from time to time, so I’m sure he’ll get to it here, shortly. In the meantime, please feel free to look through the various topics here, and don’t hesitate to weigh-in on any that catch your fancy.

> http://www.talkigy.com/

Hey, that’s cool.  I like new ideas, and get tired of seeing the same old stuff all the time.  What this reminds me of are dynamic webpages, automated documents, self-writing books, and natural language generation.  I’ve been looking at the inverse, deconstructing webpages, documents, and books into automated knowledgebases.  In theory, this should be a two-way process.  If a chatbot can understand them, then it should also be able to create webpages, documents, and books.

One of my ideas is that a chatbot should be able to “ingest” a novel, for instance, and then be able to converse through the personality and circumstances of any character in that novel.  In other words, you should be able to converse with any character in that book through the chatbot.  And ideally, in the case there is insufficient data within the book, it should be able to go out and fill in any gaps in the background with outside information or research.

Could you describe the general architecture of talkigy.com, without giving away too many of your trade secrets?  ;^)

[Actually “New to Chatbots.org” is a pretty lame topic header….  Maybe Dave Morton will get inspired to change the topic header to something like “David Lovatt’s new talkigy.com”, just saying.]

lol, I’ll give it consideration, Marcus. Let’s see what David thinks of the idea, first.

Dave thanks for the welcome and please go ahead and change the thread title as you see fit.

Marcus, you clearly understood the intent of talkigy straightaway, emphasizing a means to publish content, (with less emphasis on e.g. trying to pass the turing test).

Architecture-wise, here are some key features:

* content is authored in spreadsheets (no coding required, see http://www.blog.talkigy.com/yes-or-no-for-scotland-send-your-answers.html for an example of the required format.)[/li]

* Content is simplified by using common synonyms and word stems, both for stimuli entered by the user and responses stored in the database.

* Responses are matched to stimuli following a hybrid hierarchy of algorithms , e.g. close pattern matching in the easy cases and then a ranking of matching key words in more difficult cases - e.g. when the database does not have a close match for an input.

* So far, managing and updating the content is a manual process, a key next step is to enable users to manage and update their own spreadsheets.

* All is developed in Python. A design goal is to keep things simple and fast performing.

More background on why I started this at http://www.blog.talkigy.com/the-idea-for-talkigy.html

Tx Dave, Dave, I’ll be checking the submissions after the weekend!

Still waiting to have my bots in your listing