Critical Reflection

• Roadmap

    If we have more time, like working during the summer break, which items should we consider to improve and refine?

    The first aspect that we would like to involve is to adjust the camera bridge above the typewriter and make it more suitable to the moving of platen knob of the machine, as for now the bridge might influence the working efficiency when people type. A good idea is enlarge the bridge and make it thinner or change another material for the camera bridge.

    The second thing that we could engage is to improve the text recognition system. Currently, we use Open CV for Python as the library to recognize the texts printed by the typewriter. But sometimes, it is not very accurate due to the lens focus of the camera or the CV recognition system. It would be a potential to regularize the training process and also fix the lens focus to improve the whole system more sensitive.

    The last aspect is the typewriter itself. As the machine is relatively old (more than 60 years old), some of the keys are not working very well, which lead the printed notes a bit blur sometimes. From this point of view, we would fix the ink tape or deep clean the key hammer to make the printing more clear and sharper.

• Critical Reflection

    One reflection and feedback from the users is the user-friendly design can be improved. When the sender plan to send a note to the receiver, he or she has to figure out how to load paper via the roller. Besides, how strong to type the keys is also an element to consider, as we found out that if the user type a bit light, the camera detection system will not work perfectly. When the user finish the note and about to send out, they have to find out the sending out button, which is not very obvious.

    Another feedback is the detection size limitation of the printed note. Some users may tend to type a long message rather than a one or two sentences' note. The camera detecting area is about one fourth of the letter size paper. Therefore, it would be difficult to detect a long message via the camera and the users need to move the paper twice in order to capture the whole message that they input via the typewriter. We may adjust the camera distance between the paper or change the detection area via Python to realize a larger area detection.