Tuesday, February 11, 2014

Steps in Software Project Proposal

Software engineering proposal is a document that a software developer submits to a business customer for acceptance. The proposal describes the problem to be solved and explains the resulting benefits to the customer.
What is Important in a Software Proposal
The key for a great proposal is to invent a great idea.
There is no “official template” for writing software proposals.
To sum up: Content is the key. Form is irrelevant.
Keep in mind that this is a proposal only, so you do not know that it will be accepted as such. Therefore, you do not want to waste time on formalities by following some formal descriptions/formats/templates.
The key is to diagnose the problem and propose a treatment, so to convince the customer to accept your proposal.
The customer wants to know first that you know what you are proposing to do.
They are not interested in idiosyncrasies of software engineering or programming. They assume you know that.
Only if you receive the customer’s approval, will come the issue of knowing how to do it.
The most important thing about a software engineering proposal is that the proposal is about the problem domainnot about programming. The proposal should clearly describe the motivation for the proposed work and the proposed solution along with its expected business value. The proposal should not contain any technical jargon. There are three key components of a software engineering proposal:
DIAGNOSE PROBLEM   ==>   PRESCRIBE TREATMENT   ==>   DESCRIBE PLAN OF WORK
How To Write a Software Engineering Proposal
To write a software engineering proposal, follow these steps:

  1. Problem diagnosis: describe the problem domain, describe clearly what kind of issues exist in the current practice that you would like to address.
    Be as specific as you can and provide as many details and examples as possible.
    Describe the problem domain and the problem that you’re planning to solve. People usually make a mistake of describing at a very high level the problem, too generic, and then make a huge leap and dive deep into the tiny detail of their own solution. You must make effort to bridge this gap incrementally. Start with a brief description of high-level context (few sentences or a paragraph), then describe some specific issues that you’re interested in, then provide more specific details about the sub-issues that your work will tackle.
    The best approach is to observe personally the current practice, so that you know what you are talking about. Another useful approach is to interview “domain experts,” people who are working in your target domain and who will be your potential customers. Expert opinion carries greater weight/credibility to your statements and analyses than a naive guess.
    Think of yourself as a journalist, interviewing your potential users and documenting their opinion about current problems they are facing and suggestions on how to address those problems.

  2. Proposed treatment: describe how you propose to address the diagnosed problems.
    What specific interventions will you introduce?
    What kind of metrics will you use to evaluate your success in solving the targeted problems?
    How will you know that you achieved your objective?
    Discuss the business value of your proposed solution. What will your customer and users gain from your proposed system that they are lacking now?
    Be as specific as possible in describing the envisioned benefits of your proposed solution. Provide example scenarios of how your proposed system will be used and explain how this solution is better than the current practice.

  3. Plan of work: make a convincing case that you know how to achieve the proposed goal. Step-by-step, go in details about what needs to be accomplished, how long it will take, and how it relates to other parts (independent vs. builds upon another part). You cannot know all the details yet, because you haven’t even started, but your plan should outline the main steps so that it is clear that you have a plan. Do not just copy a generic plan from a textbook, because that looks lame.
    Describe your team. What are the strengths and expertise of each team member? Explain why your team size is adequate to tackle the problem, and why the problem size requires your team and not fewer people.
    Keep in mind that this is only an initial plan so that you can give your customer a preliminary estimate of costs and expected completion date. You will need to adjust both of these estimates as you progress, but hopefully not by much.
    State how you will know that you succeeded. How will you measure the success of your system in addressing the customer’s problem that you diagnosed?
Your proposal must be written in lay language, plain English, and you should avoid any engineering terminology (unless your problem domain is an engineering process, i.e., you will develop software to improve an engineering process).
The proposal should accurately describe the user experience, though in lay language rather than using software engineering jargon. The proposal is about the user experience of the proposed system, so this must be as accurate as possible. Otherwise, the decision to accept or reject the proposal will be based on inadequate information. On the other hand, the proposal should avoid discussing the implementation details of the system (how it will be done). It is useful, though, to include what is necessary to accomplish the proposed goal, such as access to certain data (e.g., financial reports, traffic reports, etc., depending on the problem domain), other resources (e.g., sensors, devices, equipment), or expertise (e.g., statistician, security expert). It helps to know whether such resources are available and at what cost.

Why Writing a Software Proposal is Difficult
It is not that writing itself is difficult. What is difficult is learning about your problem domain, how things are done now, how your proposed system will change that. Even if you’re familiar with the domain, it takes time to think new solutions. For example, you may be using a parking garage every day, or even working as a garage operator, but it takes a great deal of time to figure out how to propose a smart parking system.
One may argue that it is to be expected at an initial stage that your proposal will be sketchy on details. Writing a detailed project proposal requires time.

That is why you should avoid wasting time on formalities and focus on what matters. Which is, diagnosing the problem and proposing a treatment. But you cannot diagnose a problem unless you have a deep understanding of your problem domain. You cannot tell your customer how you will improve their work unless you know very well how they are doing their work now!
This is where the problem arises. You may be excited about writing new programs, but you are not excited about learning biology or finance or restaurant functioning. But you cannot diagnose a problem and propose a treatment unless you know your problem domain. You must know how your customer does their business now and how your proposed system will affect your customer’s work. It is important that you are excited about developing something you feel is valuable and important.
A key quality of a great software engineer is that he or she is willing (scratch that, excited) to learn new problem domains.
A great software engineer will learn medicine if he or she is to develop a medical software system; he or she will learn sociology if he or she is to develop a social networking system (popup quiz: what’s Dunbar's numberand why Path limits your social network to 150 friends?); the engineer will go observe how restaurant personnel do their work if he or she is to develop a software system to help run a restaurant; etc.
Software Project Must Include Programming
When deciding about the project, the most important thing to keep in mind is that the software product you will develop should require lot of programming. This is a design rather than a programming course. However, in order to learn good design, the process must include actual implementation of the design. Therefore, the project must include programming.
For example, creating a data-bank and processing this data is a programming task. If your system just allows the user to set or modify values of different database records, that is not data processing. Examples of programming (data processing) include, data-bank re-organization, keyword-based search and retrieval, filtering and summarization of the data-bank, selecting a small group of users for notification, etc. However, just saying: The system will provide search facilities, may mean that you are using SQL database calls and doing no programming at all. And, programming is done in a programming language, such as C, C++, Java, or C#.
Note: Designing web pages passive web pages which contain only HTML is not programming.Therefore, your project can include web page design, but that is not enough.
On the other hand, designing active webpages using AJAX, Flash, etc., is programming and is perfectly acceptable for the class project.
Focus more on automatic data processing: not what user can manually do, but what system does automatically to save the user’s effort. For example, suppose you are designing a system for online purchasing of airline tickets. You could incorporate the following data processing features:
  • When user queries for the available flights, sort the results by price or some other attribute before displaying them;
  • Allow pending reservation requests for a price below a certain threshold, or in case another passenger cancels a completely booked flight;
  • Design a feature for marketing to the users based on their travel history;
  • If the flight gets delayed or cancelled, the system should automatically send notification to the passengers;
Strive to reduce the user’s manual interaction with the system and increase the system “smarts”! Let the system do the work for the user! Nobody is impressed with a system where the user has to invest hard work to get even simple tasks accomplished. Make every effort to reduce the number of clicks and keystrokes necessary for task completion.
Keep in mind that the key purpose of this course is that you learn how to do modular design of software and how to document the design using symbolic representations, i.e., UML diagrams. Thus, you should avoid flashy user interfaces and proprietary software. These will definitely not contribute positively to your grade, and may have negative effects. Limit your software to the tools commonly and freely available on the Web.

Choice of the Topic
The proposed topic can be almost anything; however, it is advisable to come up with something novel and innovative to keep you interested in its realization, and because innovativeness tends to make impression on the reviewers.
You could get an idea for interesting service(s) by exploring these web-sites:
Your project may or may not be Web based, but it must include programming.The process of casting out the right-size project for a given period of performance can be very frustrating. You cannot ever know what you’ll encounter while working on it. I believe that it’s better to start more ambitious and later scale down, if necessary. Start with an initial idea of desirable services and discuss as much details as possible with your partners to figure out whether or not you as a team can develop it in the given time frame. Of course, you’ll have chance to adjust your goals during the semester, as you learn more details about your target product.

3.   Working from an Existing Proposal

If you like one of the project ideas described here, your main task for the proposal is to describe how different will be your project from the projects that were done by past students in this course.
Borrowing From Past Projects
You may borrow ideas and implementation from past projects as much as you like. It is fine that you borrow 100% from a past project, take everything and then build on top of it! In fact, this approach would be preferred instead of seeing you reinvent the wheel.
However, what is important is that you contribute new value! For example, you can implement better some features/functions from an existing project; your implementation may be faster, or easier to use. Or, you may add novel features / extensions to an existing project.

4.   Proposal Format, Submission, and Feedback



Proposal Format
Each proposal should contain the following:
  1. Project title, Group number
  2. URL of your project’s web-site (for free web hosting, see here)
  3. Team profile:
    1. Individual qualifications and strengths (such as: programming, design, presentation, documentation, management and organization)
      Note: It is expected that every team member shall be involved in all project activities; this only indicates individual strengths, not their sole responsibilities
    2. Name of the elected team leader, if any (having a team leader is optional)
      The team leader should act as facilitator, to organize group meetings and generally keep track of project activities
  4. Proposed project description:
    Regardless of whether you are proposing a completely new project or building on one of the example software projects, your project description must provide detailed explanation of your proposed project,
    You may also describe typical customers for your proposed system and mention if you already have a customer or someone interested in your proposed project.
  5. Product ownership:
    Split your team into pairs of students (a team of 6 students will have 3 pairs), and each pair should list what specifically they will contribute to the end product:
    1. Functionality: which functions, if any, you will contribute, such as customer registration, data capture and storage, data processing to extract statistical parameters, etc.
    2. Qualitative property, if any, that you will contribute, such as tune up the system performance to achieve response time under x seconds, or develop and evaluate an easy-to-use user interface for this-and-this specific functional feature, or ensure confidentiatlity of a specific set of data, etc.
    Product ownership is critical to demonstrate that each team member will play a clearly defined role in the proposed project.
Proposals that are missing any of the above sections will be returned without review.
Proposal Submission
Because the project is a team work, each team submits a single document for the entire team. Email a PDF document of your proposal to both the instructor and the TA on or before the due date given here.
Submission deadline: 5:00 p.m. on the due date.
Here is a PDF writer you can use from Windows. Macs automatically convert postscript to PDF by clicking on the file icon. There are UNIX utilities that convert postscript to PDF for Linux users; make sure you include all fonts in the created PDF.
In case of novel project proposals (not based on the projects described in the class lecture notes), email immediately the breakdown of the individual contributions to the proposal.
Feedback
We will provide written feedback only for novel project proposals via email in case we feel the proposal should be revised. No feedback will be provided for proposals based on the projects described in the class lecture notes. The only two reasons for revision could be:
  • Project does not appear to include programming
  • Project does not appear to be sufficiently ambitious

The students are encouraged to be ambitious in the first iteration of the project. You will have chance to revise your project objectives at the end of the first iteration,