We have fallen into a steady workflow with our documentation, and this week was spent completing it. As of now, we are two sections away from finishing. With the pace of the work winding down, I've had some time to explore Rochester. I went to see Niagra Falls yesterday, which was amazing, and also took the bus to a few malls in the area. I must say I've enjoyed the research experience overall.
This week we came to the realization that the scope we defined for the project was too ambitious. It has become clear that we will be unable to go any further with implementation than we have. With this in mind, we have changed our focus to consolodating our findings in a research document for future researchers to utilize. We also decided to make a final round of audio files and another survey to test the conclusions we had come to based on the last set.
We spent this week documenting our progress thus far. Specifically, we documented the survey and audio file generation processes. After this, we began to refine our related works document, which details similar research conducted previously, and which we began during the second research week.
This week has been a fairly slow one. We received some feedback from our expert programmer and spent most of the week following up with him and making new audio files. Towards the end of the week we reviewed a parsing tool we had discovered earlier, and it was determined that due to the format of its output, we should probably explore the possibility of using xml instead. How we will tag source files with xml without first parsing them eludes me, but I will look into that more next week.
Week 5, though shorter than last week due to 4th of July, was no less productive. With our mini project completed, and both of us a bit more comfortable with python and event driven programming, we have once again focused our attention on how we might modify the screen reader. Rather than look at how we might implement our modifications however, we decided to first verify the merit of those modifications. We looked for a free audio editor and decided on one called Audacity, which we used to add custom audio cues to the screen reader reading code snippets. We had generated 5 distinct audio files, each with unique audio cues (and one with none at all) by the end of the week. We also utilized surveygizmo.com to produce an online survey to test the effectiveness of our audio files. The survey has been brought to the attention of an expert blind programmer, who will be the only participant until we have refined the files further. I must say I'm quite eager to hear what he thought of the files.
This week has been amazing! On Monday we switched our focus from looking at source code to familiarizing ourselves further with python. In this vein, we decided to collaborate on a small project of our own, using wxpython, a python framework for wxwidgets. It has been really interesting working with guis, and I'm particularly happy that we're working with wxwidgets, as I should be able leverage the understanding I gain here when I return to C++ programming in the fall. Furthermore, the process of event driven programming has been demystified entirely for me after finding some great resources to help me understand it. Towards the end of the week we were wrapping up our little project (which was an address book) and my mentor lent me some books on object oriented design and design patterns. Overall, I would say this has been my most enjoyable week thus far; while we haven't made any headway on our main project, I've learnt more than I have during any other week this summer.
This week we began to seriously look at the NVDA source code. We spent the majority of the week engaged in browsing the source, and subsequently downloading an executable version to our computers. Initially we were looking through it online, however we ultimately made use of Bazaar (version control software) to create our own branches, with the intention of modifying the code at a later date. This was the first time I've actually used version control software, and it was a great primer for me to learn more about it, which was quite interesting. We spent the last two days of the week downloading and installing the dependencies necessary to build our branch of the code. Looking through the source code for the screen reader has put into perspective how little python I know. While I familiarized myself with the basics last week (functions, classes, control structures, built in data structures), the nuances of the language are still new to me, and I often found myself having to look things up as I went. In addition, the screen reader makes heavy use of event driven programming, with which I am quite unfamiliar. I will likely have to revisit my Python learning resources next week, if I am to extract meaningful information from the NVDA source.
My second week at RIT has been somewhat slower than the first. I've gotten my dorm in order and fallen into a sort of routine of work, errand, and leisure time now. I learnt this week that my colleague is actually from the same college as I am (Morehouse), which is interesting. As far as the project goes, we've thrown ourselves completely into the screen reader now, and are no longer involved with the LEGO robots at all. The details of the project are starting to coalesce, and become more and more clear in terms of things to be accomplished. We spent the end of last week, and the beginning of this one, reading research papers on sonification to get a better understanding of the problem space. This week, we spent a good deal of time documenting that research, and also making the first steps toward applying what we are learning to an actual screen reader. Rather than develop a screen reader from scratch, as I suggested in the first post, it seems we will be customizing an existing, open source screen reader called NVDA for this project. In this vein, we are familiarizing ourselves with Python (in which NVDA is written) so that we can better navigate the source code. I have found this to be an extremely rewarding endeavor thus far, and hope it continues to be so as summer progresses.
This has been an interesting first week. I got settled into my dorm and explored the campus and surrounding area a bit. I also met my coworkers and was introduced to the project they had been working on. They have been working with programmable LEGO robots to develop activities to introduce blind and visually impaired students to programming. I got to see some of the stuff they had done, and me and the new interns actually worked through one of the activities. Later in the week my colleague and I (it seems the two of us will be working on a separate project) were told what we would be working on in the weeks to come. The goal of our project is to develop a screen reader to help the visually impaired read source code. I've never done anything similar, so I'm sure it will be quite interesting and illuminating work.