This unit considers the equivalent of the old MYP Design cycle. At university level for a Computer Science course, the design cycle is known as the SDLC: System development life cycle (some books substitute “system” with “software”).
The major sections of the SDLC are:
We will examine each section for what needs to occur.
Aim of phase 1: Figure out what the project needs to accomplish.
There are typically 4 elements to this:
Find out exactly what it is you have to do
A scope document allows all parties to agree in writing as to what is and is not included within the project. It effectively becomes the contract and establishes common expectations for client and creator.
Scoping should be SMART
Ensure your scopes identify time, financial and resource constraints!
It is inevitable that a project scope will change over time. When that occurs, beware the challenges of scope creep vs scope discovery:
Possible strategies for obtaining requirements from stakeholders
Failure to involve all relevant stakeholders may lead to software that is not suitable for its intended use! (The manager doesn’t necessarily always know what the clerical staff do!)
Effective collaboration and communication between all parties: client, developer, end users.
Consultation should occur continually throughout the lifecycle to identify problems early.
Be aware of privacy issues – being able to get honest, frank information from a stakeholder without fear of retribution (eg: a staff member who might have valuable insight into how some part of the system doesn’t work the way management thinks it does) – create an environment where you can extract those valuable nuggets
Split off into pairs - one taking the role of customer, one taking the role of developer. Customer comes up with an app/software project they want the developer to “design”. Developer must ask questions to ascertain:
Reverse roles when ready.
Save these project outlines as we will continue to use them.
When researching your project, don’t forget to account for international factors.
Does your program only have to work for a homogeneous group of people in the same location? Or might you have to deal with different timezones, different languages, different conventions in date formats? Things can get very complex very quickly as these videos demonstrate.
The Problem with Time & Timezones - Computerphile
Internationalis(z)ing Code - Computerphile
Your requirements are generally split into functional and non-functional requirements.
Aim of phase 2: Create a plan on paper before hacking away at code
There are typically 3 elements to this:
There are a range of diagrams that may be useful when designing a new software project. We will learn a few key ones now…
Dataflow diagrams (DFDs) come in two flavours:
Symbols for DFDs
An example of the two diagrams for a lemonade stand might look like:
Level 0 diagram
Rules for DFD
|Context diagram||Level 0 diagram|
|Only 1 process (it represents your entire system)||Identifies the key processes within your system (typically 3 to 8)|
|Identifies all external entities your system interacts with||Identifies all external entities your system interacts with|
|Dataflow arrows should identify if the external entity is an input, output or both||Dataflow arrows from entities must go to/from a process. You never want external entities reading/writing direct to your data sources|
|Arrows do not need to be labelled||Data flow arrows should be labelled indicating what data is travelling across that path|
|No data store elements should be identified||Data stores may or may not be required dependent on the project|
Let’s look through a few simple examples.
Restaurant food ordering system
Supermarket customer app
Stock market trading app
For one of the projects dreamed up in class during phase 1, create the context and level 0 DFD.
For example, here is a level 0 DFD from a student dreamed up “Party Finder” app.
Not to be confused with the logic flow chart.
The main difference between system flowchart and program flowchart is that a system flowchart represents an entire system while a program flowchart represents a single program.
Software development is a complex task. It is not possible to write programs for the entire system directly. Therefore, it is necessary to model the system to get a better understanding of the system. Furthermore, there are different diagrams that help to understand the functionality of the system. One such diagram is a flowchart. It is a diagrammatic representation that illustrates a solution model to a given problem. System flowchart and program flowchart are two types of flowcharts.
A system flowchart is a diagram that describes how an entire system operates. It helps to recognize the flow of operations in the system. It also helps in preparing the required documents of the system.
Provides a top-down approach to explain how the different parts of the program are put together.
For our purposes we can treat wireframes, mockups and prototypes as generally referring to the same thing (which they do just to different levels of detail). In this sense, we are referring to sketches or diagrams that protray screen shots of your intended product.
Purpose? Demonstrates the proposed system to the client.
Through building multiple prototypes along the way, and iteratively reviewing your planning documents, it is hoped you can avoid this kind of scenario…
Some cliche prototypes
An example set of wireframe diagrams follows. You can see that it would make things very clear to the client what is being proposed!
Search for “free wireframe tools” and you will find a bunch of useful tools you can use to help you with this.
Additional recommended reading:
(not in IB syllabus unless doing Databases option)
If you expect them to pay your bill, make sure they’ve agreed to everything… in writing.
You had the signed and agreed scope, now have them agree to the design you’ve built as the proposed incarnation of that project before you start serious development!
Aim of phase 3: Build it compliant to the design
Some issues that need to be considered in this phase:
Will your software be available for purchase as:
A lot of major products are moving to a SaaS model (eg: Microsoft Office 365, Evernote, Adobe Suite). Sometimes you aren’t even installing anything on your computer because you are using the “software” through their web portal. Eg: Google Docs.
Modern tools available give wider options for how your new technology might be implemented. Specifically a huge development that is comparatively recent is the emergence of cloud computing.
Amazon Web Services, Google Cloud, and Microsoft Azure are the three big global players in this space at the moment, though there are others.
What technologies will you use to build your project and why?
Common concerns/questions about using the cloud reside around you lacking ultimate control over where your data resides, who has access to it. Trust in the cloud provider is necessary for this approach to work.
Elements to evaluate if weighing a cloud-based project:
Degrees of cloud based computing.
Issues regarding the maintaining compatiblity of your customers data as they migrate to your new system.
How are you going to deploy your tool?
How will you manage updates/patches to your tool?
The lesson of Google Chrome in relation to auto-updates done “right”. There are a multitude of reasons why Google Chrome burst onto the browser market to rapidly dominate market share. See Full Circle Design if you are interested in some well articulated reasons. The aspect that is of relevance here is automatic updates. No longer does the user have to be prompted (nagged) to download updates, go through an install screen, possibly reboot their computer etc; Google Chrome keeps itself up to date for you behind the scenes. This meant new features, bug fixes, security patches, are all rolled out automatically.
What implementation / conversion method will you use?
What can you see as the plus/minus/interesting of each?
See what you can come up with, then read this
For more detail including examples, read changeover techniques @ smallbusiness.chron.
Elements to be considered within the testing phase:
The quality of your user documentation will affect the rate of implementation, or even overall success, of the new system.
What to provide?
What is best for YOUR users in YOUR project? No one-size-fits-all approach
Along with the documents, there might be a need for actual “training” too…
Data can be lost through a variety of ways. How will your system protect from:
The amount of effort/expense dedicated to data loss prevention should be judged by the answer to: How would the worst case scenario affect my business?
Eg: hotels/airlines losing all record of reservations? Banks losing record of all account balances? loss of medical records?!
A data loss prevention reigeme should include a mix of: failover systems, redundancy, removable media, offsite/online storage
3-2-1 rule of backups:
Other backup related considerations
Does the cloud count as a backup?
So, your project is up and running. What’s left to do? A post implementation review!
Gather stakeholder feedback on the whole experience (design, migration, training) not just the functionality of the final product.
What improvements or enhancements do they suggest?
Start the cycle all over again with your identified changes!