List of features ready to be reviewed - Demo(V1)

[tesYolan]

Going to our first major demo, appropriately listing all the things that are done and ready for consumption so that we could see how far or near we are to a first release.

So what are the status of:

• [ ] Ghost Integration
• [ ] Speech to Text integration
• [ ] Basic hardware requirement specification for speech recognition to run on raspberry pi. Something akin to recommended software to capture audio from the environment.
• [ ] Text To Speech integration
• [ ] Basic hardware requirements for decent audio output.
• [ ] Cross-compile toolchain stability tests. Something similar to continuous integration tools that lets evaluate validity of chain.
• [ ] Benchmarks for timing tests for different visual, audio and nlp libraries incorporated in this software package.
• [ ] Also check on button startup to run the demo.
• [ ] Check on button deploy on a empty memory card. Do we have an image people can download and run on their setup.
• [ ] Make our images publicly available for easy of consumption by various people across the glob.
• [ ] List the softwares that are installed on the our image. Most people may need to inquire about what packages that we have installed in our toolchain. So listing all packages that are available in the images we provide.

For now, this are the things that are on top of my mind. In time we would see if there are additional things to be added in our first demo.

[smigad]

Ghost is integrated, @Eskender-B is writing the rules and stories. Speech to text is using PocketShinx and if the mic has good quality and the speaker has a good american accent, it's works fine. The problem is when there's noise and when the speaker's accent isn't as intended. We need to add a small noise subtraction module in there for the first problem and a grammar analysis of what was heard for the second one so that the robot can say "I'm sorry can you repeat that please" or something like that when gibberish is heard. Audio output is not a problem, however we need to worry about input and get good mics. Cross-compiler has issues now, it's going to be changed entirely. In the mean time, there's always an updated deb package here about an image, there's a very minimal debian we slimmed down for the pi with only the basic deps of opencog and this repo... should release that soon.