A Multivocal Mapping Study on Artifact Traceability Complexities in

Practice

Zaki Pauzi

and Andrea Capiluppi

University of Groningen, The Netherlands

{a.z.bin.mohamad.pauzi, a.capiluppi}@rug.nl

Keywords:

Software Traceability, Multivocal Mapping Study, Evidence-Based Software Engineering.

Abstract:

Artifact traceability is essential for managing the relationships between artifacts produced during the soft-

ware development lifecycle, yet achieving effective traceability in practice remains a complex challenge. This

study explores the multifaceted nature of traceability in real-world settings, providing actionable insights for

researchers, practitioners, and tool developers aiming to enhance traceability practices, improve software qual-

ity, and support project success. Drawing from 56 academic papers and 15 grey literature sources, this study

synthesises ﬁndings from scholarly research, industry reports, practitioner experiences, and expert opinions.

Key challenges include the lack of standardised processes and tools, difﬁculties in maintaining traceability

over time, balancing automation with human involvement, and fostering effective stakeholder communication

and collaboration. Two critical open challenges emerge: achieving semantic interoperability and managing

scalability in complex systems. To address these, we recommend targeted efforts towards standardisation and

the development of incremental, adaptive techniques for traceability management.

1 INTRODUCTION

Multivocal mapping studies (MMSs) in software en-

gineering research offer a unique approach to syn-

thesising diverse perspectives and insights on com-

plex topics within the software engineering ﬁeld. Un-

like traditional systematic mapping studies that fo-

cus primarily on academic sources, MMSs incorpo-

rate a wide range of voices, including academic liter-

ature, industry reports, practitioner experiences, and

expert interviews (Garousi et al., 2016; Neto et al.,

2019). This was particularly chosen given that syn-

thesising the data including sources from grey lit-

erature (e.g., white papers, technical documentation,

blog posts) (Lefebvre et al., 2008) has immense value

as they provide timely, context-speciﬁc, and diverse

insights that complement academic research. This

supports evidence-based decision-making in industry

settings, given that our scope is focused on “in prac-

tice” (Garousi et al., 2016; Garousi et al., 2019) – we

are focused on experiences and reports. As we look

into the complexities of artifact traceability in prac-

tice, we considered sources outside of academic liter-

ature, albeit through a systematic process of scoping

https://orcid.org/0000-0003-4032-4766

https://orcid.org/0000-0001-9469-6050

with inclusion/exclusion criteria to reﬂect high qual-

ity data that is in scope for our study. Solely focus-

ing on scientiﬁc research will miss out on alternative

perspectives and diverse voices from industry practi-

tioners, consultants, and tool vendors. These are not

typically published in academic settings.

The following research questions were outlined

based on existing research and artifact traceability in

practice, and will be assessed as part of the MMS:

RQ1: What are the demographics of reviewed lit-

erature?

Rationale: This information gives us an overview of

the metadata of our sources. This is particularly im-

portant to better understand the impact and quality of

our papers in scope.

RQ2: What are the reported key complexities in

artifact traceability in practice?

Rationale: Through collating these, we are able to

consolidate pain points and challenges. This allows

us to understand the perils and pitfalls of artifact

traceability in practice, so we can beneﬁt researchers

and practitioners alike in identifying these.

RQ3: What are the pertaining existing chal-

lenges?

Rationale: From the key complexities identiﬁed, we

collate the themes and denote these as pertaining

(open) challenges. This provides a collection of areas

754

Pauzi, Z. and Capiluppi, A.

A Multivocal Mapping Study on Artifact Traceability Complexities in Practice.

DOI: 10.5220/0013458400003928

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 20th International Conference on Evaluation of Novel Approaches to Software Engineering (ENASE 2025), pages 754-761

ISBN: 978-989-758-742-9; ISSN: 2184-4895

to focus on for actionable insights to address them,

which are not in scope of this paper for now.

This paper aims to tackle these questions by con-

ducting a thoroughly focused, yet comprehensive,

multivocal mapping study. The scope of our study

speciﬁcally focuses on complexities and in practice –

relating to the difﬁculties of achieving traceability in

practice.

2 BACKGROUND

Artifact traceability in practice refers to the ability

to systematically document and track the relation-

ships and dependencies between various artifacts cre-

ated throughout the software development life cycle

(SDLC). These artifacts can include requirements, de-

sign documents, code modules, test cases, and more.

In practice, this involves establishing and maintaining

links between these artifacts to ensure that changes

made to one are reﬂected appropriately in others.

Secondary studies looking into traceability of soft-

ware artifacts are sporadic across different domains,

with varying directions and focus in terms of reported

ﬁndings and recommendations. For example, spe-

ciﬁc to requirements engineering (Tufail et al., 2017;

Wang et al., 2018; Lyu et al., 2023; Saleem and Min-

has, 2018) and focusing on machine learning applica-

tions (Pauzi and Capiluppi, 2023; Aung et al., 2020).

To the best of our knowledge, there is not yet a study

that includes grey literature to look into reported key

complexities in artifact traceability. Our contribution

to this area is much needed to formulate what needs to

be focused on to address the key challenges pertaining

artifact traceability in practice.

3 METHODOLOGY

For our MMS, we followed the published guide-

lines for conducting multivocal reviews by Garousi et

al. (Garousi et al., 2019), namely in the following:

1. Search strategy and source selection (including

study quality assessment)

2. Data extraction and synthesis

3. Report results (based on RQs)

Based on these steps, we present Figure 1, which

shows the overview of our MMS methodology.

3.1 Search Strategy and Planning

We extracted the content and metadata of each piece

of literature using a systematic approach and ap-

plied various tools to gather all publications necessary

within our scope. For the academic literature, multi-

ple library databases were used, such as ACM Dig-

ital Library and Scopus. For the grey literature, we

used the same search string as the academic library

database search, except a more manual approach had

to be done (further reported in this section). This

planning was done to ensure comprehensiveness in

the study; to address the research questions at hand.

Threats to the validity of our study strategy will be

discussed in Section 5.

3.2 Search String

Table 1 shows the terms relevant to our search and

their synonyms. These were derived to expand the

boundaries of semantic keywords that are relevant to

the research topics. We have separated the terms ac-

cording to the relevant theme it belongs to, and only

the most relevant synonyms (to our research ques-

tions) are shown in Table 1.

For any search strategy, the construction of the

string is necessary as is enables transparency for val-

idation and reproducibility. This search string is used

for library database searches (further explained in this

section) and used in web search engines for non-

academic sources. An effective search strategy is usu-

ally iterative and beneﬁts from trial searches using

various combinations of search terms derived from

the research question(s) (Kitchenham and Charters,

2007).

3.2.1 Trial of Potential Candidate Terms

Synonym terms are then evaluated through a robust

process. Figure 2 shows the combination of terms that

were tested. We grouped the synonyms according to

common properties they share, denoted by the ovals.

Each of these groups are then evaluated on effective-

ness through trials and a decision is then made. Green

coloured groups were those chosen.

3.2.2 Decision and Final String Output

• Theme 1: (top-down order) Main term and types

of artifacts.

• Theme 2: (top-down order) Main term and off-

shoot terms.

• Theme 3: (top-down order) Main terms, types of

complexities.

• Theme 4: (top-down order) Main term, synonym

term (less common), synonym term (more com-

mon), parent term.

A Multivocal Mapping Study on Artifact Traceability Complexities in Practice

755

Figure 1: Filtering of academic literature based on selection process.

Table 1: Terms table.

Theme Term Synonyms

Software artifact software artifact source code, tests, documentation,

requirements

Traceability traceability trace link recovery, trace retrieval

Complexities complexity difﬁculty, obstacle, barrier

In practice in practice real-world, industry, experience

Figure 2: Grouped synonym terms: potential candidates for search string.

We speciﬁed the following search string (in order)

to extract all related publications within our scope:

(”software artifact” OR ”software artefact”) AND

(”traceability” OR ”trace link”) AND (”complex-

ity” OR ”difﬁculty”) AND (”in practice” OR ”in-

dustry”)

3.3 Source Selection – Inclusion and

Exclusion Criteria

To ensure our results are reﬂective of recent research,

we have imposed inclusion criteria in terms of pe-

riod scope: years 2014 to 2023 inclusive. Spanning

a period of a decade (ten years) in consideration, we

aim to ﬁll in the gap of studies that predated our start

year and focus on more recent complexities of arti-

fact traceability in practice. For exclusion, we have

disregarded content that is unrelated to (software en-

gineering) traceability, such as other reviews and non-

complexities reported.

For the exclusion criteria, we used the following

ﬁlters to ﬁlter out the papers that are not within our

scope:

1. Duplicates: repeated entries

2. Language: non-English papers

3. Data: incomplete (missing) data

4. Reviews: other reviews, surveys, and mapping

studies

5. Context: irrelevance to our deﬁned research topics

3.4 Data Extraction for Peer Reviewed

Literature

Table 2 shows the literature databases that were used

for our ﬁrst step in data extraction. The aim was to

gather all relevant publications related to our study

topics by using the search string deﬁned. The extrac-

ENASE 2025 - 20th International Conference on Evaluation of Novel Approaches to Software Engineering

756

tion was done either by exporting from the web page

(via manual extraction using the Web UI) or the API.

Table 2: Details of index databases used.

Database Extraction type Results

Scopus API 3

Google Scholar API 872

Springer Link Web UI 51

ACM Digital Library Web UI 35

IEEE Xplore Web UI 8

WorldCat Web UI 4

Total count 973

Google Scholar was ﬁrst used to get the most re-

sults possible: despite the abundance of false pos-

itives, it has the potential to considerably extend

the outreach of the systematic search (Harzing and

van der Wal, 2008). What we observed was that

using Google Scholar was enough to capture more

than 95% of other results we obtained from the other

databases. Regardless, we expanded our search be-

yond the search engine to ensure the comprehensive-

ness of our search strategy.

After the cleaning step instrumented by the exclu-

sion criteria, we gathered a total of 56 papers held

by libraries worldwide. We have also ensured that all

these were peer-reviewed publications. These were

extracted, along with the metadata, and compiled into

a spreadsheet consisting of all the information and

content for each paper. The full list of all selected

papers in scope can be found online

3.5 Data Extraction for Grey Literature

Beyond academic literature, we also expanded our

search to ofﬁcial technical documentation, white pa-

pers, and case studies published by companies and

reputable institutions. We speciﬁcally chose these

outlets as our inclusion criteria for grey literature

(ﬁrst tier literature), following the quality assessment

checklist for grey literature (Garousi et al., 2019).

However, we make an exception for blog posts (orig-

inally categorised as the lowest tier in the guidelines)

that are authored by the organisation themselves pub-

lishing about their products, as these sources do fulﬁl

the criteria for a Tier 1 source. Given that our scope is

targeted to “in practice”, we evaluated sources of re-

ports and conclude that only those published ofﬁcially

will be included in our study – this is based on the cri-

teria of authority and outlet type, which complements

our academic sources.

On top of the inclusion and exclusion criteria

listed above, we added the following exclusion cri-

https://github.com/zakipauzi/enase-2025/blob/main/

papers.csv

teria for technical documentation and white papers:

1. Tool/platform does not present or mention trace-

ability.

2. Tool/platform does not support end-to-end trace-

ability.

3. Unofﬁcial documentation (not authored by an of-

ﬁcial afﬁliate or endorsed).

For case studies, the following exclusion criteria

were added:

1. Post/report not authored by ofﬁcial afﬁliate or en-

dorsed.

2. Case study does not address fully or part thereof

artifact traceability.

We ensured that these do not impede on our pur-

suit of comprehensive literature sourcing by including

grey literature sources, while simultaneously ensuring

high-quality reviewing by conforming to these crite-

ria. The complete list of grey literature in the scope of

our study is shown online

, due to space constraints.

4 RESULTS

4.1 RQ1: Demographics

As part of the impact and quality analysis, we look

into the sources and publishers of our academic liter-

ature. Our pie chart

shows the distribution of papers

selected in scope for our study. The majority of our

papers are from journals and conference proceedings.

We have also included students’ theses that have made

the selection criteria.

For citation count per year, we can see 3 out-

liers in our box plot

– these are top cited publi-

cations per year, corresponding to the papers (Guo

et al., 2017; Mahmoud and Niu, 2014; Abbas et al.,

2022b). Correction has been made on one of the pa-

pers recently (Abbas et al., 2022a). Despite the ci-

tation count to be, arguably, a weak indicator of re-

search quality for some (Aksnes et al., 2019), for the

purpose of our study, we consider citation count as a

factor in research impact, and we will analyse these

further in Section 5.

https://github.com/zakipauzi/enase-2025/blob/main/

gl source.csv

https://github.com/zakipauzi/enase-2025/blob/main/

rq1 pie.png

https://github.com/zakipauzi/enase-2025/blob/main/

rq1 box.png

A Multivocal Mapping Study on Artifact Traceability Complexities in Practice

757

4.2 RQ2: Key Complexities

Based on the academic literature in scope, we have

identiﬁed and grouped together four key complexities

to effective artifact traceability in practice. We chose

to only present the most common of them where each

of these have been present in the papers (for sim-

pliﬁcation purposes). Although there are also cases

where papers cover more than one complexity, we

only present the main complexity that is the most rel-

evant for each paper. Due to space constraints, we

have uploaded the mappings between complexities to

the papers online

1. Lack of standardised processes and tools

2. Difﬁculty in maintaining traceability over time

3. Trade-offs between automation and human in-

volvement

4. The need for effective communication and collab-

oration among stakeholders

4.3 RQ3: Open Challenges

Based on the key complexities identiﬁed in the pre-

vious section, we have also uncovered the following

open challenges that organisations struggle with: se-

mantic interoperability and scalability in complexity.

Achieving consistent and meaningful links be-

tween artifacts across diverse tools and domains re-

mains a challenge due to differences in terminology,

evolving artifacts, and varied stakeholder perspec-

tives. Overcoming this challenge requires standardis-

ation efforts, integrated frameworks, and automation

techniques to ensure seamless communication and in-

terpretation of traceability information. Managing

traceability becomes increasingly difﬁcult in large-

scale software projects due to the sheer volume of

artifacts, varying levels of granularity, and dynamic

nature of software development. To address this chal-

lenge, efﬁcient storage and retrieval mechanisms, in-

tuitive visualisation tools, and adaptive traceability

techniques are needed to cope with the complexities

and scale of traceability information effectively.

5 DISCUSSION

In this section, we discuss the results of our MMS and

the threats to the validity of our study.

https://github.com/zakipauzi/enase-2025/blob/main/

mapping complexities.csv

5.1 RQ1: Demographics

As illustrated in our pie chart

, the distribution of

publication types reveals that the majority of selected

papers are sourced from journal articles and confer-

ence proceedings. This ﬁnding underscores the sig-

niﬁcance of academic research in shaping our under-

standing of artifact traceability, with peer-reviewed

journals and conference venues serving as primary

outlets. Additionally, the inclusion of student theses

and grey literature meeting our selection criteria high-

lights the diverse range of sources considered in our

study, particularly as they report on case studies and

tools. The analysis of citation counts per year further

contributes to our understanding of the impact and in-

ﬂuence of publications within the domain of artifact

traceability. It is notable that three outliers (Guo et al.,

2017; Mahmoud and Niu, 2014; Abbas et al., 2022b)

emerge as the top-cited publications per year. In all of

these papers, the authors employed machine learning

in the semantic representation of artifacts to automate

traceability.

5.2 RQ2: Key Complexities

The key issue of standardised processes and tools is

reported as one of the key hindrances to effective

artifact traceability in practice. Without established

guidelines and uniform methodologies, organisations

struggle to maintain consistency and synchrony in

traceability practices across different stages of the

software development life cycle (SDLC). The lack of

standardisation often results in ad-hoc approaches to

trace link creation, leading to inconsistencies, errors,

and inefﬁciencies in traceability management. Conse-

quently, stakeholders face difﬁculties in tracking and

managing trace links, impeding their ability to accu-

rately understand relationships between artifacts and

make informed decisions based on traceability infor-

mation.

The difﬁculty of maintaining traceability over

time commonly relates to the increasing complexi-

ties of ever-evolving artifacts and their dependencies.

Establishing trace links is commonly focused more

during the early stages of the life cycle, and in some

cases, visualisation tools are used to represent these

traces. The issue becomes compounded when these

traces are not maintained, and it becomes laborious

and difﬁcult to ensure traceability is updated. Most

of the papers tagged to this issue propose tools and

solutions to address this difﬁculty, although the chal-

lenge that comes with this does not necessarily dis-

https://github.com/zakipauzi/enase-2025/blob/main/

rq1 pie.png

ENASE 2025 - 20th International Conference on Evaluation of Novel Approaches to Software Engineering

758

appear. What used to be a manual task, traceability

solutions using automated techniques with machine

learning have taken the limelight in recent publica-

tions. The key issue with this, however, is the compro-

mise between automation and human involvement.

The distribution of papers relevant to each iden-

tiﬁed issue is fairly equal with the exception of the

fourth key issue: The need for effective communica-

tion and collaboration among stakeholders, which is

also the main reported issue for all the grey litera-

ture in our scope. We have also observed that there

are multiple instances where these issues overlap and

lead to one another.

In answering RQ2, we had to cherry-pick the key

complexities surrounding artifact traceability in prac-

tice, which is typically not the main focus of these

papers, and sometimes can be obscured in the text.

Nonetheless, these four that were identiﬁed enabled

us to map to and derive existing open challenges that

were considered to be persisting in practice. Figure 3

shows the mapping of these links.

5.3 RQ3: Open Challenges

5.3.1 Semantic Interoperability

In the context of artifact traceability, achieving

semantic interoperability involves establishing and

maintaining meaningful links between artifacts across

various tools, platforms, and domains involved in

the software development process. One of the key

complexities contributing to the challenge of seman-

tic interoperability is the heterogeneity of tools and

data models used in software development. Differ-

ent teams and organisations often employ a variety of

tools for requirements management, version control,

issue tracking, and testing, each with its own termi-

nology and data structures. This diversity makes it

challenging to establish meaningful connections be-

tween artifacts, as the same concept may be repre-

sented differently across different tools.

To address the challenge of semantic interoper-

ability, efforts are needed in several areas, such as the

following:

• Standardisation: Developing standardised ontolo-

gies, vocabularies, and data models that can be

shared and reused across tools and domains to fa-

cilitate consistent interpretation and exchange of

traceability information.

• Integration and Middleware: Building integra-

tion frameworks or middleware layers that enable

seamless communication and data exchange be-

tween heterogeneous tools and systems, abstract-

ing away the underlying differences in data for-

mats and structures.

• Automation and Machine Learning: Leveraging

automation techniques, such as natural language

processing (NLP) and machine learning, to au-

tomatically infer and maintain traceability links

based on textual, structural, and semantic sim-

ilarities between artifacts (Pauzi and Capiluppi,

2023).

5.3.2 Scalability in Complexity

Scaling with complexities is not unique to artifact

traceability, yet it remains as an open challenge that

organisations have to handle daily. The sheer vol-

ume and dynamic nature of traceability information

in large-scale software projects is a major contribut-

ing factor. As software systems grow in size and

complexity, the number of artifacts, relationships be-

tween artifacts, and traceability links increases, pos-

ing signiﬁcant challenges in managing, querying, and

visualising traceability information effectively. More-

over, traceability information may need to be cap-

tured at various levels of granularity, from high-level

requirements to low-level code elements. Managing

traceability at different levels of abstraction and detail

while preserving meaningful relationships between

artifacts adds to the complexity of traceability man-

agement. This is also compounded with continuous

changes, updates, and revisions throughout the devel-

opment life cycle. To address scalability in complex-

ity, we recommend innovative approaches and tech-

nologies that can do the following:

• Visualisation and Exploration: Develop intuitive

visualisation techniques and exploration tools that

enable stakeholders to navigate and analyse com-

plex traceability networks, identify dependencies,

and gain insights into the relationships between

artifacts.

• Incremental and Adaptive Techniques: As trace-

ability is ever evolving, so does the need for main-

taining and managing the links. By focusing on

techniques and tools that allow incremental and

adaptive methods to manage traceability, we re-

duce the burden of tracing complexities as they

evolve real-time. Smaller and more frequent trace

link recoveries are much simpler to handle and

maintain overtime.

At the backdrop of these challenges that were

identiﬁed, there is some overlap with the grand chal-

lenge in traceability that was published more than a

decade ago: making traceability ubiquitous in soft-

ware and systems (Gotel et al., 2012).

A Multivocal Mapping Study on Artifact Traceability Complexities in Practice

759

Figure 3: Mapping of key complexities to open challenges.

5.4 Threats to Validity of MMS

In this section, we outline the threats to validity iden-

tiﬁed throughout our mapping study process. Based

on a recent map of threats to validity in systematic lit-

erature reviews in software engineering (Zhou et al.,

2016), we looked into all possible similar threats that

would emerge from conducting our MMS. The fol-

lowing are some key threats identiﬁed:

Construct validity – The chosen inclusion and ex-

clusion criteria for evaluating literature may not ac-

curately capture the nuances of complexities in arti-

fact traceability in practice, leading to biased results.

This is particularly true for published reports and case

studies, given that the majority of published articles

are biased towards successes and improvements. Re-

gardless, we focused heavily on these challenges that

may not be explicit in the literature; this is done with

a thorough analysis and synthesis of available infor-

mation.

Internal validity – The literature selected for our

study may not be representative of the entire body of

research on artifact traceability complexities, poten-

tially skewing the conclusions drawn from the avail-

able evidence. By introducing grey literature, we ex-

pand the search scope to beyond academic literature,

which is necessary given that our focus is “in prac-

tice”.

External validity – The papers in scope of our

MMS may not be representative of the broader pop-

ulation, limiting the generalisability of the ﬁndings.

Comprehensiveness of search is pivotal to address this

threat, and this is why we used a search aggregate en-

gine for literature that indexes multiple databases. Al-

though more work is needed to be done to remove the

false positives, we wanted to ensure that our ﬁndings

can be generalisable.

6 CONCLUSION

In this paper, we conducted a multivocal mapping

study (MMS) to explore the complexities of artifact

traceability in software engineering practice. Our

study addressed three key research questions: (1) the

demographics of reviewed literature, (2) reported key

complexities in artifact traceability, and (3) existing

challenges pertaining to traceability.

Regarding the demographics of the reviewed lit-

erature, our analysis revealed that the majority of se-

lected papers are from journal articles and conference

proceedings, with a notable inclusion of student the-

ses meeting our selection criteria. Furthermore, an

examination of citation counts per year highlighted

several top-cited publications. In terms of key com-

plexities, our study identiﬁed several common chal-

lenges faced in achieving effective artifact traceabil-

ity in practice. These include the lack of standardised

processes and tools, difﬁculties in maintaining trace-

ability over time, trade-offs between automation and

human involvement, and the importance of effective

communication and collaboration among stakehold-

ers.

Our exploration of open challenges revealed two

signiﬁcant areas of concern: semantic interoperabil-

ity and scalability in complexity. These challenges

underscore the need for standardisation efforts, in-

tegrated frameworks, and automation techniques to

address semantic inconsistencies and manage trace-

ability at scale effectively. Overall, our ﬁndings pro-

vide valuable insights into the current state of artifact

traceability in software engineering practice and high-

ENASE 2025 - 20th International Conference on Evaluation of Novel Approaches to Software Engineering

760

light areas for further research and improvement.

6.1 Future Work

While our MMS has provided a comprehensive

overview of the current landscape of artifact traceabil-

ity, there are several avenues for future research and

exploration in this area. First, future studies could

delve deeper into speciﬁc industries or domains to

understand how traceability challenges vary across

different contexts. Additionally, further longitudinal

studies could investigate the evolution of traceabil-

ity practices over time and assess the effectiveness of

interventions and tools in addressing identiﬁed chal-

lenges. It is clear that there is an imminent need

for continued research and development of innovative

tools and techniques to support artifact traceability in

practice. This includes the exploration of automated

tracing algorithms, integration of traceability mecha-

nisms into existing development workﬂows, and the

development of frameworks for assessing the quality

and completeness of traceability information.

REFERENCES

Abbas, M., Ferrari, A., Shatnawi, A., Enoiu, E., Saadat-

mand, M., and Sundmark, D. (2022a). Correction to:

On the relationship between similar requirements and

similar software: A case study in the railway domain.

Requirements Engineering, 27(3):399–399.

Abbas, M., Ferrari, A., Shatnawi, A., Enoiu, E., Saadat-

mand, M., and Sundmark, D. (2022b). On the re-

lationship between similar requirements and similar

software: A case study in the railway domain. Re-

quirements Engineering.

Aksnes, D. W., Langfeldt, L., and Wouters, P. (2019). Ci-

tations, citation indicators, and research quality: An

overview of basic concepts and theories. SAGE Open,

9(1):2158244019829575.

Aung, T. W. W., Huo, H., and Sui, Y. (2020). A literature re-

view of automatic traceability links recovery for soft-

ware change impact analysis. In Proceedings of the

28th International Conference on Program Compre-

hension, ICPC ’20, page 14–24, New York, NY, USA.

Association for Computing Machinery.

Garousi, V., Felderer, M., and M

antyl

a, M. V. (2016). The

need for multivocal literature reviews in software en-

gineering: complementing systematic literature re-

views with grey literature. In Proceedings of the 20th

International Conference on Evaluation and Assess-

ment in Software Engineering, EASE ’16, New York,

NY, USA. Association for Computing Machinery.

Garousi, V., Felderer, M., and M

antyl

a, M. V. (2019).

Guidelines for including grey literature and conduct-

ing multivocal literature reviews in software engineer-

ing. Information and Software Technology, 106:101–

121.

Gotel, O., Cleland-Huang, J., Hayes, J. H., Zisman, A.,

Egyed, A., Gr

unbacher, P., Dekhtyar, A., Antoniol,

G., and Maletic, J. (2012). The Grand Challenge of

Traceability (v1.0), pages 343–409. Springer London,

London.

Guo, J., Cheng, J., and Cleland-Huang, J. (2017). Semanti-

cally enhanced software traceability using deep learn-

ing techniques. In 2017 IEEE/ACM 39th International

Conference on Software Engineering (ICSE), pages 3–

14.

Harzing, A. W. K. and van der Wal, R. (2008). Google

scholar as a new source for citation analysis. Ethics

Sci. Environ. Polit., 8:61–73.

Kitchenham, B. A. and Charters, S. (2007). Guidelines for

performing systematic literature reviews in software

engineering. Technical Report EBSE 2007-001, Keele

University and Durham University Joint Report.

Lefebvre, C., Manheimer, E., and Glanville, J. (2008).

Searching for studies. Cochrane handbook for system-

atic reviews of interventions: Cochrane book series,

pages 95–150.

Lyu, Y., Cho, H., Jung, P., and Lee, S. (2023). A systematic

literature review of issue-based requirement traceabil-

ity. Ieee Access, 11:13334–13348.

Mahmoud, A. and Niu, N. (2014). On the role of seman-

tics in automated requirements tracing. Requirements

Engineering, 20(3):281–300.

Neto, G. T. G., Santos, W. B., Endo, P. T., and Fagundes,

R. A. (2019). Multivocal literature reviews in soft-

ware engineering: Preliminary ﬁndings from a tertiary

study. In 2019 ACM/IEEE International Symposium

on Empirical Software Engineering and Measurement

(ESEM), pages 1–6.

Pauzi, Z. and Capiluppi, A. (2023). Applications of natural

language processing in software traceability: A sys-

tematic mapping study. Journal of Systems and Soft-

ware, 198:111616.

Saleem, M. and Minhas, N. M. (2018). Information retrieval

based requirement traceability recovery approaches-a

systematic literature review. University of Sindh Jour-

nal of Information and Communication Technology,

2(4):180–188.

Tufail, H., Masood, M. F., Zeb, B., Azam, F., and Anwar,

M. W. (2017). A systematic review of requirement

traceability techniques and tools. In 2017 2nd inter-

national conference on system reliability and safety

(ICSRS), pages 450–454. IEEE.

Wang, B., Peng, R., Li, Y., Lai, H., and Wang, Z. (2018).

Requirements traceability technologies and technol-

ogy transfer decision support: A systematic review.

Journal of Systems and Software, 146:59–79.

Zhou, X., Jin, Y., Zhang, H., Li, S., and Huang, X. (2016).

A map of threats to validity of systematic literature

reviews in software engineering. In 2016 23rd Asia-

Paciﬁc Software Engineering Conference (APSEC),

pages 153–160, Hamilton, New Zealand. IEEE.

A Multivocal Mapping Study on Artifact Traceability Complexities in Practice

761