Modified on by VijaySankar
Note: This is the fourth post in our series of Managing Your Requirements 101. Read the first three posts here:
What is traceability? Or more specifically what is requirements traceability? Well rather than repeat what is already a good collection of definitions, I’ll refer you to http://en.wikipedia.org/wiki/Requirements_traceability
. From there I’d summarize three elements to requirements traceability:
Following the life of a requirement – from idea to implementation
How requirements impact each other, and how requirements impact other development lifecycle artifacts (such as designs, tests, tasks, source code, hardware specs, etc.) and vice versa.
The decomposition of requirements – from high level user/customer/market needs to system, sub-system, software or hardware component requirements; and transformation into design specifications and the implementation realization of the requirement.
Traceability in this context is about relationships between requirements at the same or different levels of detail, and between requirements and other lifecycle artifacts as listed above. It also extends to relationships beyond those directly involving requirements – i.e. the relationship of a defect report to a test case – this is referred to as ‘lifecycle traceability’. Traceability relationships can be of multiple types, for example:
Satisfaction: a system requirement (or more likely a number of system requirements) ‘satisfies’ a user requirement e.g. system requirement ‘The engine shall have at least 200bhp’ satisfies user requirement ‘The car shall be capable of accelerating from 0-60mph in under 8 seconds’.
Verification: a test case ‘verifies’ a requirement e.g. test case ‘0-60mph acceleration test’ (consisting of a number of test steps) verifies user requirement ‘The car shall be capable of accelerating from 0-60mph in under 8 seconds’.
Dependency (often used where interfaces are concerned): a requirement ‘depends’ on another requirement e.g. requirement ‘the power socket shall take 3 pins’ depends on requirement ‘the plug shall have 3 pins’.
Basic traceability establishes a relationship or link between one or more elements. Typed traceability adds the relationship type with its associated semantics (examples above). Rich traceability (ref: Requirements Engineering, Hull, Jackson & Dick, Springer, 2004) adds additional information on the traceability relationship such as the rationale explaining why a group of systems requirements satisfies a particular user requirements; or as is often the case, you can’t be 100% certain on specification or design decisions, you might document any assumptions you made in deriving a set of systems requirements from a user requirement. The rich traceability approach is particularly valuable in heavily regulated industries and safety-critical systems where audit trails of decisions made are vitally important to provide assurance and reduce risks.
Once traceability has been established there are multiple ways in which it can be viewed and reported on. Perhaps the oldest and most commonly recognized method is the traceability matrix where you can see the intersection between two sets of requirements and a check or cross shows where a link exists. This method doesn’t scale particularly well since the matrix could become very large. It’s also sometimes used for creating the links, but it’s not ideal for that either since you can typically only see a small amount of information on the requirements.
Another way to see traceability is to pick a starting point, e.g. the user requirements and display the related systems requirements alongside the user requirement they are linked to, in a traceability column. You can typically choose how much detail of the linked requirement is displayed, and you can even make it recursive, going down as many levels of requirements as you need/is practical to manage in a single view.
Graphical displays are great for getting a bigger picture view of traceability rather than immediately focusing in on the details of particular relationship. You can explore the traceability tree, zooming in/out or collapsing/expanding parts of the tree, or changing the focus (starting point) of the tree.
But what about in agile development, I hear you cry? Well that could be another topic in its own right – watch this space - but relationships still exist between typical artifacts created in agile approaches (such as between product features and user stories), and I argue that as long as traceability is created ‘as you go’ and automated by tools as much as is practical, that it’s even more essential to stay informed when changes are happening rapidly and ensure you are looking at the correct versions of related artifacts.
In a follow-on post in this Requirements 101 series, I’ll take a look at what traceability can be used for – highlighting where its application can bring significant value to your projects. But for now I’ll leave you with a few resources below that I’d recommend you take a look at, and ask you to let me know if you think this post was useful (or not!) and provide any feedback or additional information using the comment function.
Requirements come in a variety of forms and level of details. Well defined and thought out requirements have demonstrated time and time again to reduce development time, increase quality and lower the cost of a software project. Projects that lack specific requirements quickly fall into many of the typical traps of project failure.
In this era of agile, mobile and ever faster-paced world of today, some teams question whether managing requirements are still relevant today. Well I say teams still need to know what they are supposed to do. The developer may have ideas but in order to deliver something useful there must be a consensus on functionality and needs. And they need more than just a general idea scribbled on a sticky note, but specifically exactly what does the business and customer expect. Developers should not be expected to fill in the gaps with assumptions. Clear definitions of requirements actually accelerate the development process by reducing uncertainty and wasteful work.
Another point of view argues that requirements only have to be defined once at the start of a project , and not worry about them again during development. Sure you could consider that approach but then we quickly realize that change is all around us. There is no way we can ever realistically say the requirements are "done". Requirements are continually evolving even as development is underway and therefore the team needs to stay attuned and make adjustments along the way. Otherwise the risk is that the end result although matching up to the original requests, fails to meet the business need upon delivery. The only way to address this effectively is to realize that requirements are an intricate part of the lifecycle that continually evolve with the business and that the delivery team needs to be in sync to deliver these expected results.
To ensure requirements are written for the entire lifecycle it is important to realize that business needs come in many different forms, formats and language. Bringing them all together in a single place, removing redundancy, and connecting interrelated content is the first step to requirements definition and management. Capturing your stakeholder and user requirements is only the first step in realizing your project. As your project progresses and evolves, these initial requirements give rise to other requirements which elaborate, satisfy, and verify them. The result is a web of interdependent artifacts and relationships which represent the nature of the dependencies between these artifacts. Navigating and understanding this web of information is critical to the success of your project. Each stakeholder in a project may be interested in different information (e.g. defect backlog, current tasks, or release schedule). What helps here is a customizable dashboard displaying all the information stakeholders, developers and testers need at any time. As a new project is started and progresses forward many of requirements you define and capture will change, be added to, or sometimes removed through the course of development. You need the capability to track requirements information changes so that you can understand what has taken place throughout that process.
To learn about automated solutions for managing requirements, I recommend listening to this informative webcast Three Reasons to Throw Away your Requirements Documents
In summary I want to remind business analysts and the rest of the software delivery teams, that in this fast-paced delivery environment today, requirements are more important than ever. They provide the roadmap of exactly what is expected of the team by the business and customers. Everyone on the team needs to appreciate that requirements are never final and that they naturally evolve with the market and business. Teams that can capture evolving requirements and implement them in a rapid and iterative fashion will find a competitive edge in the market. Automating requirements provides the additional key benefit of traceability and collaboration to identify gaps and the impact of changes while facilitating the engagement of experts on the team to improve the quality of results.
Modified on by VijaySankar
Just like the famous (mis)quote of Mark Twain, rumors of the demise of the requirements management tool IBM Rational DOORS are not only exaggerations but in fact untrue. I’d like to put straight here some of the myths and misunderstanding that I’ve heard and seen perpetuated:
Myth: IBM is discontinuing support and development for the DOORS 9.x series
Truth: IBM continues to develop the DOORS 9.x series with the same level of development resources. We have new functionality to announce in 2013 and plan further releases in the years ahead. In addition IBM has recently invested additional development resources in creating a new Requirements Management tool called IBM Rational DOORS Next Generation (DOORS NG).
Myth: IBM is replacing the DOORS 9.x series with DOORS Next Generation (DOORS NG) and expects customers to migrate now
Truth: DOORS 9.x will be developed in parallel with DOORS NG for many years to come. DOORS NG was released for the first time in November 2012. The tool has many functions DOORS 9.5 does not have, e.g. fully functional web client, type and data re-use (requirements & attributes), team collaboration, task management, but also does not have all of the capabilities relied on by DOORS 9.x users. We encourage users to evaluate the capabilities of DOORS NG and start projects or move projects to DOORS NG when practical and beneficial to the project.
Myth: To move to DOORS Next Generation, I’ll need to buy a whole new tool
Truth: Customers with active support & subscription on their DOORS 9.x licenses are entitled to use DOORS Next Generation - it is included as part of the DOORS 9.x package. Customers can choose to use either DOORS 9.x, DOORS Next Generation or a combination of both. There is no need for additional purchases or even for an exchange of licenses. All licenses are available to customers in the IBM license key center.
Myth: IBM offer no migration path from DOORS 9.x to DOORS NG
Truth: IBM do offer functions to transition into using DOORS NG for existing DOORS 9.x customers:
Work with DOORS 9.x and DOORS NG alongside each other - both tools support linking of information between both databases.
Work with suppliers by exchanging requirements through the standard exchange format of ReqIF. This works best between DOORS 9.x and DOORS NG but is also designed to work with other RM tools.
Where DOORS NG projects wish to be initialized with data from DOORS 9.x we offer re-factoring functions to harmonize the transition of data from DOORS 9.x to DOORS NG.
Myth: Everyone knows IBM’s strategy on requirements management and DOORS
Truth: IBM takes great care to regularly communicate our product strategy and roadmap at trade shows, IBM conferences and on webcasts. If in doubt, come and ask IBM! Please submit questions using the comment feature here or contact your IBM representative.
For information about the products, visit -
IBM Rational DOORS
IBM Rational DOORS Next Generation
On Thursday 14 March I presented on an IBM sponsored Dr Dobbs webcast on the topic of ‘3 Reasons to Throw Away Your Requirements Documents’. If you didn’t catch the webcast, you can view it on-demand and download the slides. If you did attend I hope you enjoyed it and in this blog post I want to answer some of the questions that I didn't get time to cover in the webcast, so scroll down to see if I've covered your question here. But first, for those of you who didn't make it, here's a quick recap of points I made during the webcast:
- What do I mean by ‘Throw Away Your Requirements Documents’? I’m not saying do away with your requirements or your requirements management process. What am I saying is invest in improving your requirements management process and support those process improvements with the right tooling. In the webcast audience poll, 60% said they are using documents and/or spreadsheets for requirements, so I set out to provide the case for why they should consider moving to a requirements management tool.
- Reason #1 – Collaboration. Using documents and spreadsheets, you can get into all sorts of problems like working from the wrong version of requirements. An integrated requirements management tool (one that has a collaborative repository that has open interfaces to share requirements data with design, test and development environments) helps you ensure that all engineering / development team members and their stakeholders are working from the right version and view of requirements for their role. I shared information from a case study on MBDA Missile Systems where collaboration was a major challenge.
- Reason #2 – Traceability. Traceability helps you get context and an audit trail for decisions made, perform coverage analysis, detect gold plating and perform impact analysis (See blog posts ‘What is Traceability?’ and ‘The uses and value of traceability’ for more explanation). But using documents and spreadsheets to create and manage traceability gives it a bad name – it becomes a tedious, error-prone overhead activity. With an integrated requirements management tool traceability links can be easily created and navigated. Traceability becomes part of the process rather than a bulk catch-up exercise, and through open interfaces it becomes easy to extend traceability from requirements to artifacts in other tools such as designs, work items and test cases. And most importantly there are automated views and reports that enable you to make active use of traceability for the purposes I mentioned above. I shared information from a case study on Invensys Rail Dimetronic where traceability is essential.
- Reason #3 – Agility. As many engineering and development organizations are looking to improve time to market and reduce costs, agile approaches are becoming increasingly popular (in the webcast audience poll 56% said they were using some sort of hybrid waterfall/iterative approach and 20% were agile), and with that comes changes to the way requirements might have been traditionally defined and most importantly to the way that changes to requirements are managed. Change is allowed and encouraged but you must have reviews and impact analysis to make informed decisions about whether to accept the change and implement in the current iteration or sprint, plan it into a future iteration or reject the change altogether. And to do that requires more than just a backlog managed in a spreadsheet or an agile planning tool. At IBM Rational we keep a prioritized backlog of user stories and epics as work items in plans managed in Rational Team Concert. User stories and epics are then decomposed and more fully described by requirements and supporting visual and textual artifacts created and managed in Rational DOORS Next Generation.
- And I started and concluded by looking at the most compelling reason for improving your requirements management process with the support of the right tooling: Return on Investment. I shared information from a case study on Emerging Health, Montefiore IT where they a achieved, among other benefits, a 69% reduction in the cost of quality (test preparation, testing and rework) within 6 months of deploying an improved requirements management process supported by IBM Rational DOORS. There’s also a follow-up video to this case study that was recorded after the client had adopted agile development practices.
Ok now onto some of the questions that were submitted but I didn’t time to answer during the webcast. I’ve divided them into sections based on the type of question so you can easily scan down to topics you’re interested in:
Alternative approaches to managing requirements
Q. What's the advantage of a requirements management tool if I could link my change management tickets to fine grained requirements (use cases, storyboards) that are maintained on a wiki or other collaboration tool (e.g. Sharepoint or IBM Connections)? It would seem that the work items would give you the needed traceability and metrics?
A. While you might be able to create the level of traceability you require, how would you report on it? Would you have to build bespoke views and reports? And would the use of a wiki for ‘fine grained requirements’ provide you with a view of requirements in context with one another as opposed to individual wiki pages? Where would you document additional properties of requirements? Could you easily reuse requirements across projects? A requirements management tool like IBM Rational DOORS Next Generation provides built-in traceability views and reports, enables to you to structure requirements in context with another in document-like views, provides user-defined properties for recording additional information and facilitates reuse of requirements.
Requirements Management and Agile Development
Q. We are mostly using water fall model, but looking into checking if Agile works for our customers. I am very interested in the role of a Requirements Analyst in the Agile world. In our world, the analyst is a facilitator of requirements gathering and not a SME on the business application we are gathering requirements for.
A. That facilitation role and the analysis skills of the Requirements or Business Analyst are still essential in Agile development. A common mistake when moving to moving to more agile approaches is appointing a ‘Product Owner’, who is a subject matter expert in the business domain you are building an application or product for, and expecting them to write the requirements or ‘user stories’. While they are experts in the business and you need that expertise on hand for Agile to work, they are not usually skilled in getting to the root goals or needs of the business problems you’re looking to address with the product or application. Without the skills of the Requirements or Business Analyst, it’s all too easy for the user stories to become about automating the way things are done today rather than addressing the real business issues in an optimal way. You can read more about the role of the analyst in Agile development in an interview with Mary Gorman of EBG Consulting.
Q. How does the IBM requirements toolset compare to more "agile" focused toolsets like Atlassian, Rally, etc.?
A. IBM is very committed to supporting agile development, particularly when agile is scaled to large, distributed development teams. At the heart of our agile development capabilities is Rational Team Concert which supports agile planning, task management and change management. As stated in the webcast though, we’ve found in our own continuous delivery process and heard from clients, that a place is needed to capture more details of requirements and their associated properties, in context with one another, together with supporting artifacts like storyboards, workflow scenarios and use cases. The integration of Rational Team Concert with Rational DOORS Next Generation provides those additional capabilities while preserving traceability to user stories and epics managed in the product backlog.
IBM Requirements Management solutions
Q. It seems this discussion is on Rational DOORS. Is Rational Requisite Pro still offered? If so then how do these two products compare?
A. IBM continues to support, maintain and respond to enhancement requests for Requisite Pro, but our future direction for requirements management for IT application development lies with Rational Requirements Composer and we provide migration support and a trade-up program. Please contact your IBM representative for more details. If you are using Requisite Pro for requirements management for complex products or embedded systems development, then you might also want to look at whether Rational DOORS or Rational DOORS Next Generation is the right move forward for your organization. But don’t worry we’re not forcing you to migrate today.
Q. You've mentioned DOORS Next Generation in your presentation, what is that? Does it replace DOORS?
A. IBM Rational DOORS Next Generation is a requirements management application on a collaborative lifecycle management platform for systems and software engineering that provides requirements collaboration, planning, reuse and lifecycle traceability. DOORS Next Generation (DOORS NG) was introduced in 2012 to take advantage of the common, collaborative ‘Jazz’ platform shared by Rational Team Concert, Rational Quality Manager, Rational Design Manager and Rational Requirements Composer; and to extend the requirements management capabilities in Requirements Composer to meet the needs of product & systems development organizations developing complex and/or embedded systems. DOORS NG will enable IBM to introduce new capabilities, faster, that would have been more difficult to deliver with existing DOORS product. However DOORS NG does not replace DOORS today. We have a very large install base of DOORS users working on programs that can last tens of years, and we will continue to support, maintain and enhance the DOORS 9.x series to meet the needs of those users. We encourage existing DOORS users to take a look at DOORS NG, to try it out on pilot projects and use the interoperability capabilities to exchange and/or link data with DOORS 9.x. Existing DOORS customers with active support & subscription are entitled to use DOORS NG without requiring an additional purchase – you can use your existing DOORS license entitlements to use with either DOORS or DOORS NG or a combination of the two. But we are not telling existing DOORS users to migrate live projects to DOORS NG today. DOORS NG in it’s early releases is attractive to organizations who don’t currently use a requirements management tool and are looking for a web based solution that also offers common platform integration with change management, agile planning, test management and design management capabilities. You can download a trial and follow development plans for DOORS NG on Jazz.net.
Q. How does Rational DOORS Next Generation compare to Rational Requirements Composer?
A. DOORS Next Generation is based on Rational Requirements Composer but extended to meet the requirements management needs of product & systems development organizations developing complex and/or embedded systems. In the first releases of DOORS NG, the web client is identical to Requirements Composer, but DOORS NG also features a rich client which is designed for usability of editing large requirements specifications. The strategy for the two products is that while Requirements Composer will be focused on the needs of business analysis and IT application development teams, DOORS NG will be focused on the needs of systems engineers and product & systems development teams.
I’d like to wrap up this post by thanking all of you who attended the webcast, participated in the polls, asked some great questions and completed the survey feedback. If you missed it, you can catch a replay or download the slides. I’d welcome any additional comments or questions here.
Today we have with us, Theresa Kratschmer writing about the importance of metrics in requirements management. Theresa Kratschmer is a senior software engineer who joined IBM T.J. Watson Research in 1996. There she worked on defect analysis, requirements, and Orthogonal Defect Classification deployment. In 2010, Theresa moved to sales where she is a technical specialist focusing on the Rational Jazz products. Previous to joining IBM, Theresa developed software for real-time, graphics, and database applications in the medical electronics industry. She can be reached out at theresak[at]us.ibm.com
Everybody is talking about metrics today. Numbers have always been important to builders, financial wizards, and politicians. If you were a software developer though, you may have taken a lot of math courses but you didn’t really use that math directly. So, why the sudden interest in software and metrics or specifically in requirements and metrics?
Metrics are important whenever you need accuracy such as building a house or sewing a dress. They are important when you need to work efficiently and you can’t afford waste. In today’s business environment, every organization out there, whether a finance company, government agency, or healthcare organization needs to work more effectively and produce more data, more projects, more software with far fewer resources. Numbers, or metrics, is the way to do this. Since requirements are the foundation of software, it is one of the best places to apply metrics. By applying measurements to metrics, we get insight into the organization’s requirements activities. We find out how much progress is being made, whether we have gaps in the downstream deliverable related to requirements and how big our project risk might be when changes occur to those requirements. Applying measurements to requirements also allows us to make continuous improvements.
So if metrics are so valuable, why aren’t more people collecting them? There are primarily three reasons
- It often takes too long to collect the metrics. By the time you end up gathering all the information you need, it’s out of date already.
- How do you interpret the information that you collect? Some people feel that if you don’t have access to a person with a PhD, you’ll never understand the numbers.
- Even if you collect the data and you correctly interpret it, is the organization actually going to do anything with that data to actually make improvements?
The best way to collect metrics, of course, is to do it automatically as the team goes through their daily activities. As they define and refine requirements, process requirement reviews, perform change management and quality management activities, data can be collected automatically, processed, and made available in real-time. This is exactly what an integrated solution like Jazz platform
does. It basically removes this obstacle and makes data collection, processing and display of data a natural part of the project team’s daily activity.
Interpretation of Data
Understanding data and resultant reports does take some skill. But it is not complicated. I always start by articulating exactly what information I want to view or what questions I need answered. Then I look for the data that will help me answer those questions. For example, you might want to which high priority requirements have no test cases associated with them. One of the capabilities with the Jazz suite that helps here is that the reports have a section called “What does this report tell me?”. When you run a report, you actually have some automatic interpretation built in. Of course, I always recommend people start simply. Your skills will grow as you get more comfortable understanding and interpreting the data. Jazz also allows you to set up filters which help you answer specific questions at the click of a button. This makes analysis automatic and very, very easy.
Improvement Actions for Continuous Improvement
Identifying and implementing actions is one of the most important aspects of ensuring continuous improvement. It’s very important when analyzing data that indicates weaknesses in your process to actually identify improvement actions. Not only do you need to identify how you will improve your project and processes you follow, but you also need to assign owners and a date for implementation to make sure you do actually make improvements.
I hope I’ve convinced you that metrics are important – well worth the time and effort it takes to collect and analyze data. By identifying and implementing improvement actions your organization can make considerable progress in working more efficiently and really achieving the ability to do more with less.
Note: This is the sixth post in our series of Managing Your Requirements 101. Read the first five posts here:
- What is requirements management and why is it important?
- How to write good requirements and types of requirements
- Why baseline your requirements?
- What is traceability?
- The uses and value of traceability
Requirements elicitation is the process of discovering and clarifying the needs, capabilities, conditions, and constraints that a project must satisfy to deliver a solution or product that meets the client/market needs. Requirements elicitation is by far the significant activity taken by a business analyst or requirements engineer. Elicitation in a traditional world begins early in the cycle – consulting or development while determining the scope and objectives of the project. Generally it extends to the analysis and design phases. We will delve in detail how this differs in agile in a later post. In either case, elicitation is an iterative and ongoing process which takes care of clarifying, refining requirements and identifying constraints and new changes. The amount of elicitation of requirements depends on where you are with the engagement lifecycle. Elicitation happens at various levels - early on to draw down the initial requirements and later on to refine them to specifics.
So what are the sources of requirements? The main sources of requirements are the stakeholders themselves. The end users, SMEs in the domain are also prominent sources for clarifying requirements. Other sources could be the regulatory requirements, existing documents, past experience and business cases.
Essentially, the requirements elicitation starts with a method adoption workshop. A Method Adoption Workshop (MAW) helps in determining the key requirements activities, templates and outputs. The workshop essentially acts as a discussion venue to determine the optimal activities for the project. Various methods can be used for these workshops. Most of them prescribe how to collect and specify requirements and in some cases how to manage them also. Some of the most used methods are SE&A Method and Custom Development Method or proprietary ones like SAP. These workshops also help in putting in place effective project change management processes. Thus effective scope management is the combination of requirements management and change management. MAW is used to tailor and adopt the project management and technical methods for the product and determine the optimal activities needed. Post MAW, stakeholders are identified and various elicitation techniques are used for determining the requirements. Techniques can vary according to the group of stakeholders. A consolidated business requirements document is created for client review.
Understanding the project environment is the key to determining the requirements. This could act as the starting point of elicitation. INCOSE Systems Engineering Handbook provides a classification of various sources of requirements - External environment (like the regulations, laws, culture, and competition), Enterprise environment (internal policies, technology), Project environment (budget, tools, project management) and Support functions in the organization. Before looking into the various elicitation techniques, some of the factors to consider are - the triple constraint - cost, schedule and scope; stakeholder influence, stakeholder access and willingness to participate, contractual deliverables and organizational experience in similar projects.
Broadly we could categorize the elicitation techniques into Structured techniques, Analytical techniques and Interactive techniques. I believe the genesis of structured techniques is from marketers. Generally these techniques include interviews, focus group discussions, surveys and workshops. Analytical techniques could include interface analysis, domain experience, and market research. Interactive techniques could include brainstorming sessions, prototyping, observation and reverse engineering. Many text books have dealt in detail these techniques and when to use them; however Unearthing Business Requirements: Elicitation Tools and Techniques by Kathleen and Rosemary deals with this topic in detail. We will look into some of the techniques in detail in a later post and also touch base of when to use which techniques.
Some useful resources
Form feeds function: The role of storyboards in requirements elicitation
Requirements Elicitation Introduction, Nancy R. Mead, Software Engineering Institute
Modified on by VijaySankar
Note: This is the fifth post in our series of Managing Your Requirements 101. Read the first four posts here:
Part 4 of this series ‘What is Traceability’ looked at the definition of requirements traceability and different types of traceability relationships. In this part, let’s look at what traceability can be used for and where it delivers value to application, system or product development.
Why is traceability necessary/important? Isn’t traceability just an overhead, an onerous documentation task that’s only done in industries where it’s mandated? In my view it’s true that requirements traceability practices have originated from industries like aerospace & defense, where one use of traceability is to show that contractual requirements have been addressed, but it also has so much more value to bring when used effectively, and you have the right tools to maintain it and report on it. The following lists some of the ways that traceability delivers value:
Context: If you can trace back from a design or test to a user requirement, you then have the reason for the existence of that design or test and through the information in the user requirement (and through any related requirements you can trace to), you have more supporting context to help create the optimal design to meet the requirement or the most effective test to verify that the requirement is met
Audit trail & compliance: Taking an example of a new person joining a project, traceability can help them navigate the project and see why particular requirements, designs, tests, etc. exist. This is also of upmost value and importance when you need to demonstrate compliance to a regulation or standard to an auditor – the traceability trail can help you quickly show you are addressing the regulation or standard.
Coverage: How do you know whether you’ve covered all the user requirements in the derived systems requirements, designs, tests, etc.? Traceability can help you here – you can see gaps indicated by where a higher level artifact doesn’t have any relationships to lower level artifacts. Of course even if a relationship exists you still need to follow it and examine the lower level artifact to ensure it does what the traceability relationship says it should, but at least you had a signpost to direct you to the right place to look.
Gold plating: As well as highlighting coverage gaps, traceability can help you investigate possible ‘gold plating’ or over-engineering. If you have a lower level element without any relationships to a higher level then you can ask the question – why does this exist if it’s not apparently satisfying a requirement? There might be a perfectly valid reason but this gives the opportunity to identify and eliminate anything that could bring unnecessary additional time and cost to the project.
Impact analysis: I think this is the most valuable reason for traceability. If you have traceability relationships in place you can following those relationships when say a user requirement changes, to identify all the related system/sub-system/component requirements, design elements, tests, work items, etc. that are potentially impacted by the change. This will enable you to fully scope out the impact of the change before it’s made (if you decide to proceed), giving you far more control over the cost and time impact of change requests. This can of course also work in reverse – if a design change is necessary, say because the original design proves infeasible, you can more easily see what impact that has if any on your ability to still meet the requirements. Both of these scenarios have great project management benefits – you can have informed discussions in the development team and with your customer/stakeholders about whether to make the change.
What’s that I hear again – what about agile? Aren’t traceability and the benefits it’s proposed to have only necessary/of value in waterfall development? Well, take another look at that list of benefit scenarios – don’t you always want to be able to do these things, regardless of development methodology? If you’re in a fast changing, evolving project don’t you need to be more informed, have the right information at your fingertips, in order that you can respond quickly but effectively? I argue that if you have the right tools in place to create and utilize traceability, that it is even more essential if you’re adopting agile practices.
Speaking of the right tools, tools can automate the various types of traceability reporting and analysis I described above. For impact analysis, tools can quickly display all of the related artifacts connected to a particular requirement. And more than that they can detect changes in a requirement that has links and automatically mark those links as suspicious. This is very effective where you have different people working on different parts of the application/system and you need to be aware of changes in other areas that might impact your work. A suspect link indicates that a change has been made to the artifact at either end of the link and that you should check that the link still holds true – i.e. if the user requirement has changed, does the linked system requirement still satisfy it? This proactive impact notification mechanism helps to avoid inconsistencies across your specifications.
So you have my views on the uses and value of traceability, but what do you think? Please use the comment function to leave feedback and additional ideas.
I’ll leave you with a couple of links that have additional discussion of the value of traceability and in particular how much traceability is enough:
This is the fifth part of our six part blog posts series on basics of requirements management. Read the remaining parts here -
1. What is requirements management and why is it important?
2. How to write good requirements and types of requirements
3. Why base line your requirements?
4. What is Traceability?
5. The uses and value of traceability
6. Revisiting Requirements Elicitation