2.744
Concept Exploration (sketch model) Reviews
Home > Assignments > Concept exploration results > Reviews for Solution Checker (works-like)


 ConceptExecution
Average Rating
 
Client 1:
Client 2:
Reviewer 3:
Reviewer 4:
Reviewer 5:
1-marginal     2-ok    3-good     4-very good    5-outstanding

Feedback on the concept explored by the sketch model

Client 1:

The software seems really well developed here. I am really impressed by the software that generates a new puzzle each time, this is definitely critical to the game and it is great that you have solved this technical issues. I think that technical design work is still needed in developing how the chocolates mechanically function in the room though. How do they slide? How do guests interact with them? How do guests move around the room? These are big questions that need to be answered to create a functioning room.

return to top of page


Client 2:

Using a camera system for this is absolutely the right way to sense this - especially since the large candy pieces will take a beating. This lets the props be 'dumb' and not need special sensors. We do this with a game that involves floor tiles and rearranging shapes.

Your puzzle generator is an awesome piece of code. That really makes it unique and interesting every time, and much more adaptive than an 'A/B' state game.

I would encourage that the 'frosting' on the chocolates be brighter colors to create more contrast to the chocolates. While what is shown on the art is far more true to reality, I think some pastels wouldn't hurt the look and would make the shapes easier to identify.

return to top of page


Reviewer 3:

It's a very technical concept to explore, so when just reviewing the storyboard, I was worried about the actual execution of the idea because I assums that it will invovles some heavy coding.

return to top of page


Reviewer 4:

Self reset is a critical component of the puzzles and for something like this (a manual puzzle), it's important to get it right.

return to top of page


Reviewer 5:

The sketch model explores an important feature of this game; what kind of sensor would make this puzzle work best. It was interesting to see that you explored RFIDs but ended up using a single camera especially because of the inherent concern that sensors could break if they are moved around too much and handled by people. It is clear that the sketch model wanted to see what design works best for the visual sensing, and it is conveyed well through depicting designs that worked well and didn't work so well.

return to top of page



Feedback on the execution of the sketch model

Client 1:

The software portion of the sketch model are very well executed, now turn your attention to the mechanical aspects.

return to top of page


Client 2:

Very well executed! Coming up with a way to make sure that the camera can see the whole room (if the puzzle really is 144 square feet - holy crud, that's huge) will be a challenge with lighting, angle, and so on.

We have been allowing the use of unistrut or truss across the center of the room to support a camera system (or in our current prototype room, a projector.) Just as long as the rig is kept small enough not to interfere with a sprinkler system.

return to top of page


Reviewer 3:

Despite of the challenge posed by the concept, I was very expressed by the execution of the model. The student was able finish a system of coding in such a short period of time, even though the model does not work perfectly at this stage, the student has already identified its drawback and I'm sure the real model will be outstanding

return to top of page


Reviewer 4:

While I think the image recognition solution is quite innovative, I wonder if it overcomplicates it a little. As you mention, some patterns are harder than others to identify. What happens when the lighting is different? or if someone is standing over the pieces for whatever reason?

return to top of page


Reviewer 5:

The sketch model is executed well because it explores which designs worked well for the visual sensing and the design that didn't work so well. On the other hand, I am interested in what method was used to visually sense and match the tiles because there may be alternative methods that can more effectively sense the current designs that you have. Are you using RGB or HSV for example to distinguish between the colors? Those two yield slightly different results especially when the colors are similar. Also, I would have liked to see a more detailed analysis on what exactly made the good design "good" so that you can use that feedback to make future designs.

return to top of page