Horizontal Traceability: Connecting the Dots for Effective Software Development

Tool support for trace model maintenance, cited in “Materials and Methods”. This paper also demonstrates how the approach implements the automatic tracing of information; a task critical for information fusion. Since the 1990s, the advantages and problems of traceability have been discussed in several academic works, including the surveys of Gotel & Finkelstein (1994) and Winkler & Pilgrim (2010) and the articles of Haouam & Meslati (2016) and Charrada et al. (2011). The first, a cornerstone paper in requirements traceability research, identified the main problems of requirements specification. The second highlighted the gap between industrial needs for traceability integrated within the development process and existing academic proposals. For further discussion of related works, the reader is referred to the next section.

• Their approach, however, is state-based, unlike that of the present study, which is event-based and focuses on the detection of dangling links.
• Since the 1990s, the advantages and problems of traceability have been discussed in several academic works, including the surveys of Gotel & Finkelstein (1994) and Winkler & Pilgrim (2010) and the articles of Haouam & Meslati (2016) and Charrada et al. (2011).
• This definition emphasizes the potential of traces in the requirements engineering domain, where traceability had its origin.
• This image presents an example to illustrate how the monitoring model helps in changing management.
• Horizontally traceable schedules support the calculation of activity and milestone dates and the identification of critical and near-critical paths.
• For more details on the process of systematic evidence generation, the reader is referred to Gutiérrez, Escalona & Mejías (2011).

This type of Matrix can be used to show you your requirements hierarchy and how higher-level work items, like Epics, decompose down into lower-level requirements, like Test Cases and Bugs. In our horizontal traceability live demos we show teams how they can build the same style of matrix in roughly 15 seconds. In this video you learn everything you need to build these matrices using simple Work Item types.

Tool support for model-driven traceability

It is therefore necessary to continue working to improve the results in order to meet challenge 3 (Ch.3. Return of investment (ROI) measurement) cited in “Materials and Methods”. These experiences, and in particular the iMedea example described in the Materials and Methods section, demonstrate how effectively the proposed approach addresses challenge Ch.4. Demonstration of the benefits of traceability in real examples managed by real users, cited in “Materials and Methods”. They plan to carry out a software experiment (López et al., 2020) based on the principles presented in Vegas (2017) to guarantee the benefits of the proposed approach in a more scientific manner. The industrial application of the approach would allow experiments to be organized within industry, offering an opportunity for further validation. The idea is to conduct controlled experiments in software companies to measure the value-added of this traceability approach for software development teams.

In practice, traceability often implies a qualitative improvement, but one which is usually difficult to measure. The difficulty of comparing software development with and without traceability management under almost identical conditions is also the main reason for the lack of systematic evaluations regarding traceability return on investment (ROI). This article presents an MDE approach supporting automated, almost “for free”, trace management in the software construction process.

Guide on how to hire software development team for your project

In fact, the Intersection Matrix  is objectively the fastest way to build relationships between two work items in Azure DevOps. The Intersection Matrix will reflect any current relationships that exist between two work items. The Intersection Matrix allows you to easily update the relationships between two work items by adding a relationship where they intersect. This image presents a part of code to illustrate how NDT-Suite implements a concrete trace. This figure presents the proceedure that a developer has to follow to apply our approach. In the example cited previously, the inclusion of the new European Health Identification Number attribute in SR-01 causes a change in both the class and the table because it requires a new attribute to be added to each of them.

When the developer creates the UIStep “Creation Anamnesis” and executes the transformation to generate test cases, the plug-in NDT-Driver asks NDT-Quality if the corresponding activity diagram is consistent and conforms to the relevant methodological principles. If it is OK, the TestUIStep “Creation Anamnesis” is created, together with a TraceLink object that stores this relationship (denominated StepTL in Fig. 7). The same process is performed several times, depending on the loops and decisions (in this example, only twice) for each activity in the activity diagram. Horizontal traceability is not required to statisfy bidirectional traceability.It identifies the realatinships among related items across work groups or product components for the purpose of avoiding potential conflicts.

Building a Work Item Type Horizontal Matrix

Horizontal traceability shows relationship among related items such as between requirements itself. Vertical traceability is a characteristic identifying the source of requirements typically from requirements to design, to the source code and to test cases. This screen presents who the tool presents the traceability matrix that is automatically generated with our approach.

The proposed metamodel is similar to several existing metamodels mentioned in the Related Work section. However, it differs in its explicit metamodeling of the traceability mechanism and the change management elements. These are represented by the metaclasses TraceRules, Change, Warning and Error (see Fig. 3).

Supporting traceability

Briand, Labiche & Yuea (2009) focus on changes between two versions of a UML model (vertical traces only) and analyze the impact of those changes using formally defined impact analysis rules (written in Object Constraint Language). Traces between model elements are not included as first citizens in their metamodel. The main approaches to traceability maintenance (Briand, Labiche & Yuea, 2009; Boronat, Carsí & Ramos, 2005; Cleland-Huang, 2006; Kassab, Ormandjieva & Daneva, 2009; Drivalos-Matragkas et al., 2010) are briefly discussed below. Laura Garcia-Borgoñon conceived and designed the experiments, performed the experiments, analyzed the data, performed the computation work, prepared figures and/or tables, authored or reviewed drafts of the paper, and approved the final draft.

The quality control plug-in verifies that each Change is performed according to the TraceRule(s); it also generates Error(s) and Warning(s). If an inconsistency is detected, a message is sent to the developer with a recommendation to re-execute or update the transformations. The methodology outlined in Section A is fully integrated into the NDT suite, a set of tools that supports development teams in the application of MDE technology and traceability. The tool and the example described above in Section C were used to validate the proposed approach. The process presented in the previous section will be applied to integrate our approach in the NDT tool suite. The approach could be integrated in the same way in other software development tools, particularly in web application development tools.

Traceability management for impact analysis

Summary schedules created by rolling up the dates and durations of lower-level elements are inherently vertically integrated. The traceability metamodel presented in the previous section is what is known in MDE terminology as a platform-independent model (PIM); that is to say, it is independent of the technology selected to develop the software. Even https://www.globalcloudteam.com/ more importantly, it is also independent of the methodology used for the software development. This means that any model-driven software modeling methodology can implement traceability, instantiating our traceability metamodel and implementing the automated generation and monitoring of traces in the tool that supports the corresponding methodology.

This section also illustrates its application in one of the real projects that were developed using the approach. Boronat, Carsí & Ramos (2005) had as their objective to provide generic traceability support to solve specific problems such as change propagation. Their metamodel provides a metaclass manipulation rule for each trace link, but change management is not included at metamodel level.

What is vertical and horizental Traceability?

Their approach, however, is state-based, unlike that of the present study, which is event-based and focuses on the detection of dangling links. The work presented by Gotel & Finkelstein (1994) is a cornerstone paper in requirements traceability research. Its authors describe the main problems of requirements traceability, including the lack of a common understanding, the need to allocate time and resources, and the gathering and maintenance of trace information.