Copy into your editor, fill in the blanks, and place in directory /pro/web/web/courses/cs190/2007/handins/1-26/LOGIN2.txt. Replace LOGIN with your CS login. Note the 2 there so that you don't overwrite the first handin! Come up with at least 3 potential projects. ================================================================ Name: Chris Hart ================================================================ Questions/comments about the syllabus or project handout: ================================================================ Project Title 1: Book Locator Brief Description This would be an online application for people to find books at any of the Brown libraries. The user would enter a query, which would be sent to Josiah to look-up the call number, and then the call number would be matched to the physical location of the book, which would be given to the user. Identify Local Users: All Brown students and faculty that use the libraries here. Assign a 1-5 score for these criteria Useful to others: 5 Has specific users: 5 Can be scaled up to PC product: 3 Divisible: 4 In C++ in Sunlab (Linux): 3 Overall: 4 Pros: Very useful in real life. It would be so much easier if you were shown the exact shelf a book was on rather than just a call number. I think the work could be divided fairly evenly. There are obvious, fairly independent parts, i.e., creating the actual map from call numbers to shelves, drawing the floorplans of libraries with the shelves, interfacing with josiah, etc. It can be easlily scaled up (include other library systems). Cons: Alot of boring busy work - figuring out exactly what call numbers are on which shelves. ================================================================ Project Title 2: Hemlock: Questions and Answers Brief Description The user types in a question, which is checked against a database of questions and answers. If the database contains the question, it will give the user the answer. If the question is not in the database, it will ask another user for the answer. When users answer they will also add keyword tags to the question. The program will learn which users are best at answering which types of questions (i.e. questions with which tag), and use that in determining who to ask which questions, making question and answer matching more and more efficient as more questions are asked and answered. Identify Local Users: Just about everyone. Anyone that uses Google or Wikipedia to try to find answers to specific questions could benifit from a working version of this with a large database. Assign a 1-5 score for these criteria Useful to others: 4 Has specific users: 3 Can be scaled up to PC product: 3 Divisible: 2 In C++ in Sunlab (Linux): 4 Overall: 4 Pros: It sounds really fun and interesting to implement. There would be alot of problem solving along the way, such as deciding how many people to ask each question, when to give up on a question because it is too hard, how are wrong answers corrected, etc. If a huge database were built by this method, it could conceivably be very useful to almost anyone. A good portion everything people use the internet for is to find answers to specific questions. Cons: It doesn't sound that easy to divide up evenly. There are alot of interdependencies among the various major things the algorithm must decide. It seems like it would take a long time to build up a useful database of questions and answers. With a class of only 10 people I doubt the database would be big enough to be useful. Every user is allowed to modify the database by answering questions, so every user must be trusted to give accurate information, which may not be a safe assumption. ================================================================ Project Title 3: Phone and Calendar Interaction Brief Description The user would call a phone number and leave a recorded message. The recorded message would be translated into text (using open-source software) and e-mailed to the user as a iCalendar file or Outlook meeting invite. Identify Local Users: Anyone that uses iCalendar or Outlook meeting invites Assign a 1-5 score for these criteria Useful to others: 3 Has specific users: 4 Can be scaled up to PC product: 3 Divisible: 3 In C++ in Sunlab (Linux): 3 Overall: 3 Pros: There's a company that would help with things like getting the VoIP phone numbers and finding the software to use to translate voice to text and answering all our questions. Cons: It seems too simple and short. I don't know how hard it is to create iCalendar files or Outlook meeting invites, but I don't think it's too hard, and there's nothing else that time consuming. We use other software for recording the call and parsing the recording into text. The bulk of the work would be mundane, technical details like sending emails and creating iCalendar files. ================================================================ Project Title 4: Brief Description Identify Local Users: Assign a 1-5 score for these criteria Useful to others: Has specific users: Can be scaled up to PC product: Divisible: In C++ in Sunlab (Linux): Overall: Pros: Cons: ================================================================ Project Title 5: Brief Description Identify Local Users: Assign a 1-5 score for these criteria Useful to others: Has specific users: Can be scaled up to PC product: Divisible: In C++ in Sunlab (Linux): Overall: Pros: Cons: ================================================================