Dave Morton - Feb 18, 2011:

I think that a leading cause of the lack of understanding we have with each others’ projects here is that there’s no really standardized nomenclature being used. Perhaps we need to get together, and create a standardized lexicon of descriptive names for various parts of our projects, so that we can get our “concepts” (normal meaning) across. Yes? No? Maybe?

Are you talking about a Wiki?

I think that a leading cause of the lack of understanding we have with each others’ projects here is that there’s no really standardized nomenclature being used. Perhaps we need to get together, and create a standardized lexicon of descriptive names for various parts of our projects, so that we can get our “concepts” (normal meaning) across. Yes? No? Maybe?

Well, as we aim to publish (at least) some research papers, one of the first steps will be to define (and describe) the vocabulary that is going to be used to describe our model and document our research.This is of course simply following proper rules (best practices?) for scientific research.

A wiki would be a good idea, sure. In fact, not only could we create some sort of standardization, and use a wiki format for it, but also, using a wiki format can expand beyond just a “list of terms” to include other useful information, too. I’ve not tried to set up wiki pages before, but I’m given to understand that they’re not hard to set up. I think I’ll set one up on my local server, here, and see what’s what.

setting up a wiki is quite easy to do. ExpressionEngine (our CMS) has a standard feature for it and it looks like this:
http://expressionengine.com/wiki

However, it’s not the technical part which is challenging, but the content part. When a wiki is not maintained, it will soon die.

Furthermore, I’d like to standarize text in a wiki in such a way that the Chatbots.org-chatbot can make use of it, and having the text maintained by Chatbots.org members. It’s not our simplest project, so therefore planned in Q4.

Moderator’s Note: Yes, I know that one of these posts is out of sequence. I missed a post, when splitting this from CR’s thread on ALEX. I can’t fix it myself, but frankly, that’s not a huge deal.

Erwin Van Lun - Feb 19, 2011:

However, it’s not the technical part which is challenging, but the content part. When a wiki is not maintained, it will soon die.

Indeed. And to have a quick start AND good maintainability afterwards, you need to build a good taxonomy first.

Having said that, if you guys want to implement a wiki then I’m happy to help. One of my specialisms is building knowledge management systems based on a wiki. I’m currently doing two corporate projects; one of the biggest maritime insurance brokers and one audio components producer that used to be a Philips subsidiary. I’m also in talks with the worldwide biggest manufacturer of carton packaging to do such a project.

Starting with an empty wiki will lead directly to the need of ‘maintainers’ who must keep the wiki in order, as the users will put in pages arbitrarily. Having a good taxonomy to start from will help a lot to alleviate this burden.

All in all, a wiki might be a great idea to maintain a database on common AI research knowledge, information and related stuff. It will at least be much more accessible then having to search the forum for all the links and information that is posted there.

Hans Peter Willems - Feb 18, 2011:

Well, as we aim to publish (at least) some research papers, one of the first steps will be to define (and describe) the vocabulary that is going to be used to describe our model and document our research.This is of course simply following proper rules (best practices?) for scientific research.

If we set up a wiki, I’ll be happy to add our vocabulary to it (when we have defined it). The vocabulary that we are developing will be matched to ‘previous art’ in the field of AI and other scientific domains that we relate to in our project (psychology, knowledge management, behavioural science, robotics, etc.).

Perhaps, we could split up the corpus in 2 parts: language/chat related and technology related. The latter probably depends on already existing ‘speech’.

Jan, I agree and I think there should be even more parts then those: speech/language is just one facet, other (from my perspective, just to name a few) are: technical implementation (coding), social implications, psychology of the mind, tactile interaction (sensors/robotics).

This is why developing a good taxonomy FIRST is so important