The project was on track and everything sailing smoothly little did we know that on 4pm the day before a storm would it. We had 90% of the project running smoothly so came in to polish off the edges. At that point setting up the raspberry pi to auto login at caused it to disable logging in my all means. Meaning we could not access anything on the pi including the code, the only way was through another lynix machine. Hours of trying emulators on our laptops had failed and resulted in rushing the 30 min journey home to find my old raspberry pi and a sd reader. With success we did recover the files and fix the log in. After that though and we went to run the code we were met with a new error, the error simple said activate the camera which we had done the week before. After some testing we found out that the camera had infect broken and we didn’t have a backup. Luckly though Sam to the rescue said he had one at his old house he thinks. And in fact he did meaning that at 11pm we were back were we had started.
When we arrived the project hand in day morning running a quick test we found out a new error. For some reason anything being sent from the pi would give the same personalty data reading even when the same text send over stander mqtt would give a mixture of readings. It turns out that somewhere in the encoding of the mqtt message something was being changed to make this happen. Unforduently we could not fix this error and would have to exslapin why it was locked into the same reading during the demo. We did provide some pre-set text on node red to show the other personalities.
Presentation
The presentation was not as polished as either of use would have liked, but when building a prototype and it failing last minute that took up our time and attention. Even though it was not as filled out as it could have been and the demo was a little bit shaky I still think we did well. Considering the setbacks the night before and the fact we took a risk with doing a personality scanning I was happy with the end result but it could have been far better.
Evaluation
I think this product could go somewhere if the technical issues could get ironed out both me and Duncan want to enter it in the ibm completion. In terms of how we worked togather I think we were a good match. Duncan is good at theory and web coding along with wood work and I can handle building the eltronices and doing the code. The combernation of our skill sets balanced the group out and I think made for a good work flow.
With the device itself, with some more work and devolvement I think we could sort out the errors with the device. For a first try and handling personality data though I think it worked very well.
When we arrived the project hand in day morning running a quick test we found out a new error. For some reason anything being sent from the pi would give the same personalty data reading even when the same text send over stander mqtt would give a mixture of readings. It turns out that somewhere in the encoding of the mqtt message something was being changed to make this happen. Unforduently we could not fix this error and would have to exslapin why it was locked into the same reading during the demo. We did provide some pre-set text on node red to show the other personalities.
Presentation
The presentation was not as polished as either of use would have liked, but when building a prototype and it failing last minute that took up our time and attention. Even though it was not as filled out as it could have been and the demo was a little bit shaky I still think we did well. Considering the setbacks the night before and the fact we took a risk with doing a personality scanning I was happy with the end result but it could have been far better.
Evaluation
I think this product could go somewhere if the technical issues could get ironed out both me and Duncan want to enter it in the ibm completion. In terms of how we worked togather I think we were a good match. Duncan is good at theory and web coding along with wood work and I can handle building the eltronices and doing the code. The combernation of our skill sets balanced the group out and I think made for a good work flow.
With the device itself, with some more work and devolvement I think we could sort out the errors with the device. For a first try and handling personality data though I think it worked very well.