COS 333 Project: Preliminary Description (Spring 2005)

Mon Jan 31 08:54:48 EST 2005


The COS 333 project is an opportunity to work on a task larger and more elaborate than any of the assignments, to design and implement a significant piece of software, working in groups of 3 or 4 people.

The intent is that this will not be just hacking, but a serious attempt to simulate some aspects of reality: choosing something suitable to work on, planning how to get it together, designing it before building it (though allowing for the inevitable changes of direction as things evolve), building it in stages, testing it thoroughly, and documenting and presenting the result, all as part of a small team. If you do it well, this should be something that you can show with pride to friends and prospective employers.

The project will involve many of the issues of software engineering as they occur in small, multi-person real-world projects. Some of this material will be discussed in class, and some will be found in assigned readings.

The considerations affecting the form of the project are:

Project Definition

A large number of real-world systems are based on what is sometimes called the "three-tier model": a user interface, a database for persistent storage, and some process(ing) between them. Many web services use this architecture. For example Amazon has a web-based user interface; the data underneath is fundamentally a large book catalog and customer information; and the process includes a wide variety of searching and retrieval operations. News and financial services are analogous: again, a user interface, a background information gathering and filing service, and mechanisms that let a client register for, access and process interesting items.

The project this year will be essentially the same as in the past few years:

This is a very open-ended project, so the big problem is likely to be defining a topic of the right size. Almost every web service will suggest something, perhaps novel or perhaps "We can do that much better"; either would likely be fine. Hiding, selecting, or merging data from existing web services might be a possibility; shopping, news and other bots are examples. Yahoo is a good place to start; one of their core competencies seems to be to invent such special services. Google and Amazon both provide program interfaces that might be a way to put a different face on some aspect of their systems.

Look around the campus for other possibilities: online maps, tours, notification services, databases, and so on are all potentially interesting and feasible (though make sure that the information that you want to use is available -- concerns for privacy, property and turf can all get in the way of a great idea). I have been talking this up with staff and faculty friends, and there may be potential projects there if anyone wants to explore a bit, and you can probably find more by talking to faculty and staff in other departments.

Some of the best projects come from noticing a place where some task is done by hand or poorly by machine when it could be done really well by a suitable program, or where something complicated could profit from a neat user interface.

In my experience, in successful projects the people are really turned on by what they are doing, whether because it's their own idea and that's sufficient motivation, or because they are doing something that someone else cares a lot about too. Either way, that kind of engagement usually leads to the best results. Be wary of projects where none of you are really interested, or where one member is driving the whole thing entirely.

Previous projects have included web-based calendars and financial systems; improvements of campus services and agencies; a variety of video rental and exchange systems (Blockbuster plus Napster); course selection and scheduling; a chat-like information service focused on the local environment; social networks. This is only a small sample, however, so give your imagination free reign.

The assignment is to create such a system, using whatever combination of existing tools and new code you like. The functionality that you should provide includes the following:

How big? There's no official requirement, but I think that it ought to ultimately involve some thousands of lines of code; hacking together a thousand lines of PHP somehow doesn't seem enough, and I'm getting to be uncomfortable with systems that are just a lot of PHP. Some systems have been as big as 10,000 lines of C++; that's more than is needed, though it was great work.

Some Options

Distributed or local? Most systems are distributed: the user interface runs on a client machine and the data is stored on a server. The processing might be at either end, or some of each. I would like to encourage systems that involve a network connection, even if it is only to a local host, so as to leave open the option of distributed operation.

Languages, tools and environment? You can use any combination that appeals for any aspect -- web-based or stand-alone; Windows or Unix or Mac or PalmPilot; Java or VB or Javascript or C#; CGI or PHP or JSP or roll your own. The only restriction here is that your running system must be readily accessible for grading, and accessible enough that you can demo it effectively in the CS building using departmental connections to the rest of the campus.

Make versus buy? Much modern software development is done by combining code and components that others have created. You can use as much of this as you like, as long as the finished product acknowledges the work of others, and has a substantial contribution of your own.

Things to Think About

I have attempted to make this as open-ended as possible. This is your chance to invent something new, or to do something better than others do. You might think of this as practice for a new e-business or e-service, the sort of e-thing that made some of your predecessors here (like Meg Whitman and Jeff Bezos and the late Phil Goldman) into e-zillionaires. Of course if you have a really good idea and sell out to Microsoft or do an IPO by the end of the semester, it's an automatic A+.

But you have only about 10 weeks, so you can't get too carried away. Part of the assignment is to plan exactly what you are going to create, what each team member will be responsible for, and what interfaces you will require between components so independently-created pieces will fit together. What schedule will you follow? How can you work on different parts in parallel and keep them integrated? How will you ensure, if your time estimates are too optimistic (as they inevitably will be), that you have a working subset, rather than a non-working collection of partially completed pieces? How will you convince a skeptical TA and instructor that you are making progress, not just writing code (or doing nothing at all)?

Since the project will involve multiple people, a major task is to divide the work into suitable pieces, with planned interfaces. Each of these components must be a separate entity that can be implemented and tested separately. But you will have to think carefully about the interfaces between them: the problems caused by poor interfaces are a recurring theme in project reports (and, happily, so are the strongly positive experiences reported by groups that did a good job on interfaces). So that each person contributes equitably, it is also necessary to be explicit about the roles of each person on the team. Each person must write a reasonable fraction of the code for the system, no matter what other role they play.

The project will represent 60-70 percent of the course grade. All members of a team will get the same grade (with the potential for a correction factor in the unlikely event of someone not carrying their fair share of the load), so you must make sure that you all agree on who will do what, by when, and to what standard.

Here are some things you should start thinking about now; this list will be augmented over the next few weeks, and we will talk about it in class as well.


The following schedule is subject to change in detail but the spirit is right. Take note.

Since the project involves more than half a semester, it is possible to develop a significant piece of software. At the same time, serious planning and steady work will be required for your project to be completed on time. To encourage planning and organization and regular activity, the project will have several deadlines that will be enforced. [Some of the dates near the end may have to be shifted a bit; there will be adequate warning.] The TAs will be responsible for primary supervision of the teams; I will act as backup and second-level management. You will be required to meet with your TA/manager each week after spring break. This is a graded component of the project, so preparation and attendance are mandatory -- no excuses.

We will also try to follow good software engineering practice as much as possible. I will lean heavily on things that have been helpful in previous offerings, such as design documents, status reports and reviews aimed at helping you to organize and me to monitor. There will also be design and code reviews involving multiple teams, where you get a chance to critique what others are doing, and have your own work assessed as well. Reviews are also mandatory and graded.

	 S  M Tu  W Th  F  S
Feb	       1  2  3  4  5
	 6  7  8  9 10 11 12	start thinking about possibilities
	13 14 15 16 17 18 19
	20 21 22 23 24 25 26	
	27 28
Mar	       1  2  3  4  5	initial project discussion  with bwk
	 6  7  8  9 10 11 12	proposal due before break
	13 14 15 16 17 18 19	(spring break)
	20 21 22 23 24 25 26	detailed project plan due
	27 28 29 30 31
Apr	                1  2	small-group design meetings
	 3  4  5  6  7  8  9	project prototype
	10 11 12 13 14 15 16
	17 18 19 20 21 22 23	alpha test
	24 25 26 27 28 29 30	beta test
May	 1  2  3  4  5  6  7	project presentations and demos
	 8  9 10 11 12 13 14	Dean's date; project due at 5 PM

  • March 4: During this week, I will meet with each team for half an hour to discuss your plans before they are set too firmly. If you need help finding a topic, please talk to me and the TAs; we're happy to respond to ideas, make random suggestions, and help steer you, starting any time. But it is your responsibility to come up with a project and partners, so don't leave this to the last minute.

  • March 11: Your team must have been formed, and you must have sketched out what you expect to accomplish in your project, and how. Each team must submit a brief (about 3 pages) design document for their project. This must include an overview of what your system will do, some background on how others attack the problem, the basic organization and components of your system, a rough schedule by week of what you expect to do, and who will be doing what. It should identify any significant risk factors and what you'll do if bad things happen. ("We're expecting a large donation to buy a server; we will drop the course if this doesn't come through in time.")

  • March 25: By this date, each group must meet with their TA to go over their project in more detail, to describe their schedule, the components and interfaces, and the allocation of people to tasks. It is highly desirable to plan a sequence of stages such that each represents a working system; if your schedule proves too optimistic, this gives you a fallback that you can still demonstrate. You must bring a draft of your design document to the meeting, and hand in a final plan by this date. Sometime in this general area, we will have meetings with 3 or 4 groups so you can discuss your own project and others in a small group.

  • April 1: Design reviews. About 3 project groups at a time will meet for an hour or so to describe their own designs and comment on others; this is a loose simulation of a peer-review of designs.

  • April 8: Prototype. By this date, you should have a bare-bones prototype that shows approximately what you are trying to do, and what your system will look like. It need not do much, but it must do some minimal part of the job. You should not be considering major or even many minor feature changes after this date.

  • April 22: Alpha test. By this date, you must demonstrate an almost working version of the core functionality of your project. "Core" means the basic operations that form the essence of your project; if you were doing Amazon-2, for instance, this might mean searching for books and accepting orders. "Almost working" means that wizards (you) can help experienced programmers (us) to use the system. Your code may crash and need restarting. But you must be able to convince us that your project can be completed by the deadline.

  • April 29: Beta test. Your code should largely work, all intended features should be installed and working, no major component should be incomplete. A determined sadist might be able to break your system, but a casual experimenter should not. Drafts of written material should be underway.

  • May 4-5: Demo days. Each group will give a 30 minute quasi-public presentation of their work. All team members must attend the presentation. You are also strongly encouraged to attend other presentations, for instruction and moral support. A one-page marketing blurb and a web page must be available for interested users. I would also like to encourage real use by naive users: try to get your classmates on other teams, your friends and roommates, and anyone else to try your system to see how well it works.

  • May 10: Dean's date. Everything must be done and handed in by 5:00 PM on this date, without exception. Final submission has to include a man page, an internals document describing the implementation, and a report on how the project and the team worked out and what was learned.


    Grading for the project will be based on a number of criteria, including

    There will be more information as we go along, to flesh out or clarify some of the sketchy parts here. There will also be class lectures on some of the pieces, like networking basics, user interface programming, and database software.

    You are encouraged to ask questions that will help clarify things for everyone. Murphy's Law applies to projects and their administration, so there will undoubtedly be screwups. I apologize for those in advance, but of course they too will be a simulation of reality...