Business Process Management Journal
Improving the quality of business process descriptions of public administrations:
Resources and research challenges
Alessio Ferrari, Hans Friedrich Witschel, Giorgio Oronzo Spagnolo, Stefania Gnesi,
Article information:
To cite this document:
Alessio Ferrari, Hans Friedrich Witschel, Giorgio Oronzo Spagnolo, Stefania Gnesi, (2018)
"Improving the quality of business process descriptions of public administrations: Resources and
research challenges", Business Process Management Journal, Vol. 24 Issue: 1, pp.49-66, https://
doi.org/10.1108/BPMJ-05-2016-0096
Permanent link to this document:
https://doi.org/10.1108/BPMJ-05-2016-0096
Downloaded on: 07 March 2018, At: 02:03 (PT)
References: this document contains references to 66 other documents.
To copy this document: permissions@emeraldinsight.com
The fulltext of this document has been downloaded 75 times since 2018*
Users who downloaded this article also downloaded:
(2018),"Understanding the nature of processes: an information-processing perspective",
Business Process Management Journal, Vol. 24 Iss 1 pp. 67-88 <a href="https://doi.org/10.1108/
BPMJ-05-2016-0102">https://doi.org/10.1108/BPMJ-05-2016-0102</a>
(2018),"A systematic approach to diagnose the current status of quality management systems
and business processes", Business Process Management Journal, Vol. 24 Iss 1 pp. 216-233 <a
href="https://doi.org/10.1108/BPMJ-12-2016-0248">https://doi.org/10.1108/BPMJ-12-2016-0248</a>
Access to this document was granted through an Emerald subscription provided by emerald-
srm:468544 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald
for Authors service information about how to choose which publication to write for and submission
guidelines are available for all. Please visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company
manages a portfolio of more than 290 journals and over 2,350 books and book series volumes, as
well as providing an extensive range of online products and additional customer resources and
services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the
Committee on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for
digital archive preservation.
*Related content and download information correct at time of download.
D
ow
nl
oa
de
d 
by
 F
A
CH
H
O
CH
SC
H
U
LE
 N
O
RD
W
ES
TS
CH
W
EI
Z 
A
t 0
2:
03
 0
7 
M
ar
ch
 2
01
8 
(P
T)
Improving the quality of business
process descriptions of public
administrations
Resources and research challenges
Alessio Ferrari
Consiglio Nazionale delle Ricerche, Pisa, Italy
Hans Friedrich Witschel
Fachhochschule Nordwestschweiz, Muttenz, Switzerland, and
Giorgio Oronzo Spagnolo and Stefania Gnesi
Consiglio Nazionale delle Ricerche, Pisa, Italy
Abstract
Purpose – Business processes (BPs) of public administrations (PAs) are often described in the form of
written procedures or operational manuals. These business process descriptions are expected to be properly
understood and applied by civil servants, who have to provide legally compliant service provisions to the
citizens. However, process descriptions in the PA are sometimes hard to read, ambiguous, or vague, leading to
false interpretations or even incorrect execution of the processes. The purpose of this paper is to focus on
improving the descriptions of BPs to be used in PAs.
Design/methodology/approach – To this end, the authors present an in-depth domain analysis, including
a literature review and interviews with PA stakeholders belonging to different realities. From this analysis,
the authors identified a set of 52 typical defects of process descriptions.
Findings – The authors provide a set of guidelines and a template to constrain the definition of these
documents and to mitigate the identified defects. Furthermore, the authors outline research challenges in the
field of quality assessment of textual process descriptions for the PA.
Research limitations/implications – This paper addresses the needs of any PA officer who has to write
an official procedure or operational manual, and should be studied by researchers who wish to provide
automatic strategies to check the quality of these documents.
Originality/value – Text quality issues have been addressed in various fields (e.g. requirements
engineering), but not in the area of BP descriptions of PAs. The contribution consists in the study of the
quality issues that occur and create problems in the practice of this particular domain. Based on this insight,
we provide directions for research that will find solutions to mitigate the issues.
Keywords Public administration, Quality assurance, Business process management,
Natural language processing, Ambiguity, Language quality assessment
Paper type Viewpoint
1. Introduction
Public administrations (PAs) are responsible for providing services to citizens in accordance
with the law. To do this, they define procedures, i.e. business processes (BPs), which civil
servants have to follow and that will result in a legally compliant service provision.
Civil servants have to learn how to execute these BPs (Di Ciccio et al., 2015); learning can be
realised via oral instructions, expert guidance, or it can be based on studying documentation,
i.e. codified knowledge of how to execute the BPs.
When studying this problem, there are two important observations to be made.
First, although graphical notations exist to capture knowledge about BPs, most PAs rely on
textual descriptions. This becomes evident when studying approaches to introduce proper
models of BPs in PAs, as in the work of Schumann et al. (2014). Indeed, graphical notations such
as Business Process Modelling and Notation (BPMN) – see Object Management Group (OMG)
(2011) – are better understood by experts than by untrained persons, and their value is higher
Business Process Management
Journal
Vol. 24 No. 1, 2018
pp. 49-66
© Emerald Publishing Limited
1463-7154
DOI 10.1108/BPMJ-05-2016-0096
Received 12 May 2016
Revised 22 August 2016
21 October 2016
22 November 2016
7 December 2016
Accepted 20 December 2016
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/1463-7154.htm
49
Quality of
business
process
descriptions
D
ow
nl
oa
de
d 
by
 F
A
CH
H
O
CH
SC
H
U
LE
 N
O
RD
W
ES
TS
CH
W
EI
Z 
A
t 0
2:
03
 0
7 
M
ar
ch
 2
01
8 
(P
T)
when complemented with textual descriptions, as noted by the studies of Ottensooser et al.
(2012), and Nawrocki et al. (2006). Moreover, graphical notations cannot model every subtlety of
a BP, and enabling them to do so would further increase their complexity and hence decrease
their understandability (Sánchez-González et al., 2011). Therefore, we argue that textual
descriptions will be used to describe BPs also in the future.
Second, using natural language always implies the risk that the message that the author
intends to express is not – or not exactly – the one that the reader understands. In PAs, unclear
instructions may result in misunderstandings, and hence either in a waste of time – when
civil servants have to understand the meaning of the instructions – or, in the worst case, in
incorrect execution of BPs. As an example, let us consider an instruction that does not clearly
define the actor who should be responsible for executing a certain task. As noted by
Sommerville (2007), such a defect can lead to several undesired consequences, e.g., nobody
might feel responsible for executing the task and, hence, the task gets delayed or not executed
at all. Such issues reduce the confidence of citizens in public services and can have severe legal
and financial consequences.
Given the need for understandable BP descriptions expressed in a clear and
unambiguous natural language that we could observe in the field, the goal of our research is
to support quality assurance of BP descriptions. Specifically, we aim at understanding the
most important defects impacting the understandability of textual BP descriptions;
providing practical and simple solutions to mitigate the defects; providing research
directions in the field of quality assessment of BP descriptions.
To identify the most important defects, we performed an in-depth domain analysis, including
a literature review and a set of interviews with PA stakeholders. The literature review revealed
that there are not previous studies on text clarity in BP descriptions in PAs. As practical and
simple solutions to mitigate the defects, we provide a set of guidelines and a template for editing
BP descriptions. In this set, we included all guidelines that we could identify from our references
and that we considered applicable to the context based on interview findings. Moreover,
to provide research directions for improving BP descriptions, we survey the current literature in
quality assessment of textual documents, we clarify the research gap and provide research
directions in readability assessment, ambiguity detection, text summarisation, and BP
model-to-text consistency. We believe that a computer-supported solution will help because
humans frequently overlook defects of descriptions when they get tired from reviewing – as has
been confirmed by research in the area of requirements engineering (Berry et al., 2012).
Our domain analysis on the quality of BP descriptions in the PA consisted of a literature
review (Section 2), and a set of interviews (Section 3). Findings from literature research and
interviews were combined to define a complete list of defects that can be found in BP
descriptions. This combination was chosen because there exists already a large body of
knowledge regarding linguistic defects in general (see next subsection), but there is
insufficient knowledge about the impact and severity of such defects, especially in our
domain of analysis, i.e. in BPs in the PA (Batini and Scannapieco, 2016). As a result, we are
enriching existing knowledge with empirical evidence from the field of PAs.
Accordingly, the rest of this paper is organised as follows. In Sections 2-4, we describe
the domain analysis performed. In Section 5, we describe the guidelines and the template for
editing BP descriptions. In Section 6, we present the research challenges that this work
opens. Here, we also discuss how previous work has addressed text quality issues in general
or in other domains. In Section 7, conclusion and future works are presented.
2. Literature review
This section discusses the most relevant works that we consulted along our literature
review, which touched the quality of a text in general (Section 2.1), and the quality of PA
documents (Section 2.2).
50
BPMJ
24,1
D
ow
nl
oa
de
d 
by
 F
A
CH
H
O
CH
SC
H
U
LE
 N
O
RD
W
ES
TS
CH
W
EI
Z 
A
t 0
2:
03
 0
7 
M
ar
ch
 2
01
8 
(P
T)
2.1 Books, associations and tools
Textual quality is relevant in many areas to achieve effective communication. In general,
it is associated with the concepts of clarity, conciseness and absence of specialised jargon.
Several books have been edited with practical recommendations on how to write using the
so-called plain English, a language that emphasise clarity and brevity. For instance, the The
Plain English Guide, edited by Cutts (1996), says, within a set of 25 guidelines on using easy
words, reducing cross references, and planning before writing, that the average sentence
length should be 15-20 words, and sentences should not exceed 40 words. Another reference
book is Style: Toward Clarity and Grace (Williams and Colomb, 1995). This book is not
structured with a set of guidelines as Cutts (1996), but is more of an education textbook,
with theoretical reflections and practical examples to be studied rather than to be consulted
as a practical manual. A peculiarity of the book is the analysis of content-related aspects
such as emphasis and elegance, which are not covered in other texts.
The use of plain English is also promoted by associations such as the Plain English
Foundation[1], the Plain English Campaign[2] or the Plain Language Association International[3].
Clarity of language also plays a large role when (technical) texts are to be prepared for an
international audience and/or for translation into other languages. In this context, guidelines
have been developed based on the concept of “International English” (Buican et al., 1993).
Some tools are available to check the quality of a text. Among them, theHemingway Editor[4],
which is free for access and download, marks with different colours sentences that are hard to
read, terms with simpler alternatives, adverbs, passive voice expressions and other quality
defects. Other tools such as the plugin for Microsoft Word 2010[5] released by the University of
Surrey are more focussed on readability checks – i.e. how easy is to read a text.
2.2 Quality guidelines in PAs
The listed books, associations and tools are not specifically targeted to PAs, but to the general
quality of a text. However, texts adopted in a PA – e.g., regulations, procedures – need to
exhibit certain quality attributes that are specific for PAs. Therefore, several PAs in the world
have defined a set of guidelines to be adopted by the civil servants writing PA-related
documents, such as regulations, procedures, press releases and even speeches. Below, we give
an overview of such guidelines.
General guidelines from the European Union (EU) are available in the document “How To
Write Clearly” (European Commission, 2012), and its recent extension “Claire’s Clear
Writing Tips” (European Commission, 2014). Besides providing recommendations on
reasoning on the expected reader of the document, the guidelines emphasise the need to use
short and simple expressions, avoid jargon and passive voice, and limit acronyms.
A detailed “Style Guide” for English, directed to editors and translators was also made
available by the European Commission (2011). The guide specifies a set of linguistic
conventions that cover punctuation, capitalisation, abbreviations, numbers and other
elements of style that an editor or translator is expected to adopt when writing/translating
official EC documents.
Also the government of the USA, within the “Plain Language” initiative[6], provides a set
of guidelines for writing PA texts (US Government, 2011), with specific suggestions on how to
structure a document. The UK government provides style guidelines in its Website
(UK Government, 2015), with a focus on writing PA documents that are published on the web.
Specific guidelines are also given on how to organise a procedure (i.e. BP) description.
Stemming from the work of Redish (2012), these guidelines recommend to describe procedures
through numbered lists and to avoid juridical jargon even in legal documents directed to
larger audience. Moreover, they recommend to keep lists of instructions short
(5-10 items), and suggest that, if lists are longer, they should be grouped into shorter lists
each with its own sub-heading. These latter guidelines are based on the empirical study of
51
Quality of
business
process
descriptions
D
ow
nl
oa
de
d 
by
 F
A
CH
H
O
CH
SC
H
U
LE
 N
O
RD
W
ES
TS
CH
W
EI
Z 
A
t 0
2:
03
 0
7 
M
ar
ch
 2
01
8 
(P
T)
Trudeau (2012), which shows that 80 per cent of the people involved in the study prefer plain
English instead of legal jargon, and that the more specialist the knowledge of the reader,
the higher the preference for plain English.
From the different sources consulted in our literature review, we selected a subset of main
sources to guide the definition of our list of quality defects of PA documents. Such sources
are the guidelines from the EU – referred in the following with the EU acronym – those of
the UK government (UK acronym), and those of the US government (USA). The selection
of these main sources was performed by the authors based on the following observations:
the texts were focussed on PA documents, and were listing defects, besides providing
solutions; the solutions listed in the other references were addressing defects already listed in
these main sources.
3. Methodology
We chose interviewees from different PA offices to have a complete view of potential
environments and problems. We selected three PA institutions: CNR-ISTI[7], a nationally
funded Italian research institution with about 300 employees (20 administrative staff ); the 2
Mears Seas Zeeën Programme (INTERREG 2MSZP[8]), a European secretariat that takes
care of managing cross-border projects within EU nations (about 20 people,
all administrative staff ); the Sportello Unico Attività Produttive (SUAP)[9], a public
Italian service for citizens, which support the procedures for opening novel businesses.
Each municipality in Italy has a SUAP office, normally with one employee in each office.
We consider the CNR-ISTI group sufficiently representative of different, large and
multi-functional PA realities since it is a large research institution, and the administrative
staff includes people with diverse expertise who have to perform different BPs at different
degrees of formalization. On the other hand, 2MSZP represents a specialised cross-national
reality, in which civil servants, who work in large and structured offices, interact with
representatives of other PAs, and provide BPs and regulations for their interlocutors in
accordance with EU laws. Finally, SUAP is a good example of a highly specialised reality,
with single civil servants acting as heads of office, with front-desk duties.
We performed two interviews (I1, I2) with the administrative staff of CNR-ISTI. Each
interview involved two civil servants that were working together in the same office.
Furthermore, we performed one interview with a EU Project Officer (I3), belonging to
2MSZP, and one interview involving two front-desk employees working at SUAP offices of
the Marche Region (I4). To have a uniform view of the gathered information, the interviewer
was always the same person – i.e. the first author. All the interviews were semi-structured
(Di Cicco-Bloom and Crabtree, 2006), i.e. based on an initial list of questions, but raising
additional questions when new and interesting aspects were mentioned. During the
interviews, we used the term procedures instead of BPs since civil servants are more
acquainted with the former term. The questions asked by the interviewer were first
concerning the types of BPs performed by the interviewees and the types of documents
describing BPs for their work (Section 4.1). Then, the discussion focussed on the types of
defects of these documents (Section 4.2).
4. Results
4.1 BPs and types of documents
The most evident aspect that we perceived from the interviews is the heterogeneity of BPs,
documents, and terminology used in the different offices. In the case of I1, the BPs
(organising committees, recruiting personnel, and managing internal practices) depend on
the circulars, which express variations on existing BPs that are not formally described
anywhere, but rely on procedural knowledge of employees. In the case of I2, BPs (redaction
of official acts related to contracts and projects) depend on internal, national and European
52
BPMJ
24,1
D
ow
nl
oa
de
d 
by
 F
A
CH
H
O
CH
SC
H
U
LE
 N
O
RD
W
ES
TS
CH
W
EI
Z 
A
t 0
2:
03
 0
7 
M
ar
ch
 2
01
8 
(P
T)
regulations, but also to a large extent on internal circulars. In the case of I3, a complex
pyramid of EU laws and regulations is used. Similarly, in I4, employees have to refer to EU
laws and regulations, but they have also to integrate them with national and local laws.
With the exception of I3, for which internal BPs are rigorously documented, in I1, I2, and I4
internal BPs are generally not formalised.
4.2 Defects of BP descriptions
The main problems of BP descriptions discussed in I1 were related to the difficulties in the
contextualisation of the content of the circulars. Indeed, in some cases such circulars appear
to refer too many external documents, and, at the same time, they miss the goal of providing
a proper summary of the content of such external documents. Such summary would help in
understanding the regulatory context of the document. Moreover, not only the regulatory,
external context is a problem, but also the practical context. Indeed, along I1, the absence of
examples on how one should put the procedure into practice was highlighted as a main
problem, together with the difficulty in finding a responsible person who knows how to
address possible issues in the implementation of the process. Finally, also ambiguity of the
BP descriptions was discussed as a relevant problem: if the content is ambiguous, the civil
servant has to take responsibility in putting the BP into practice without knowing whether
his/her choice is correct from the regulatory point of view. For example, one the interviewee
mentioned a circular that states that there should not be any conflict of interest between the
participants to a commission for the selection of new personnel. According to the
interviewed civil servant, the circular indirectly states that the director of the institution
should never participate to commissions, since these are composed by members of the
institution that might be influenced by him. However, this interpretation is not explicitly
stated anywhere in the circular, and the decision of including the director in the
commissions is left to the civil servants who organise them.
During I2, which we recall was composed of people of the same institution of I1,
the conversation took another direction. Indeed, this office has to integrate the content of the
circulars with the content of EU and national laws, and one of the first problems raised was
the overlapping and contradiction between the content of the different official documents,
which leads to practical difficulties in the implementation of BPs. A considerable amount
of time is spent in deducing what is the correct BP to implement for their tasks, given the
different documents. Another largely discussed problem in I2 was the structuring of
the different documents. Indeed, interviewees of I2 complain that it is often hard to understand
the category of a document – which can give hints on the expected impact of the document –
also because the documents are structured in a non-standardised or inconsistent way.
An interesting story about the consequences of poor document clarity was also mentioned in I2.
One of the interviewees told that the civil servants of the institution once received a circular
about a national procedure for obtaining a unique coding number for financed projects.
The circular was unclear, and was never implemented by the civil servants. After some months,
when the unique project number was required by other procedures, it emerged that the civil
servants ignored the circular because of its poor clarity. Hence, within one month, the institution
decided to deliver a specific training course on the content of the circular.
Concerning I3, for which operational manuals that describe each single activity of the
employees are constantly updated, the main problem discussed was the redundancy of these
documents. According to I3, this is not only a problem of time required to read such
documents. Indeed, documents tend to summarise the content of other laws, regulations and
manual, without necessarily being faithful to the source. Hence, the reader might have a
wrong view of the sources, and implement the BP in an incorrect way.
During I4, the main problems raised were related to the unclear motivation of both
rules and external references in the documents describing the BPs. The absence of a
53
Quality of
business
process
descriptions
D
ow
nl
oa
de
d 
by
 F
A
CH
H
O
CH
SC
H
U
LE
 N
O
RD
W
ES
TS
CH
W
EI
Z 
A
t 0
2:
03
 0
7 
M
ar
ch
 2
01
8 
(P
T)
justification leaves the civil servant – which in I4 interacts directly with citizens – in an
uncomfortable situation. S/he has to perform some activity, or ask the citizen to perform
an activity, without a clear rationale. Hence, s/he can be subject of unjustified complaints
to which s/he cannot give a reasonable answer different from the annoying expression:
“It’s the law”.
From the interview transcripts and from the literature review, we isolated 53 different
defects of BP descriptions in total. The complete list of defects is reported in Table I.
The defects are partitioned into five categories: clarity of the documents, clarity of the
sentences, synthesis, internal coherence, and external coherence – i.e. coherence with other
documents. Each defect has an identifier, which will be referred in the remainder of the
paper, when we will specify solutions to address the defects. In addition, for each defect,
the main source is reported, using the previously introduced acronyms: EU, UK and USA for
the different sources coming from the literature review, I1-4 for the interviews.
5. Guidelines and template
To provide simple means to address the defects of the BP descriptions identified in our
domain analysis, we provided two means: a set of guidelines for editing BP descriptions
(Section 5.1); a template to constrain the definition of BP descriptions (Section 5.2).
5.1 Guidelines
A set of guidelines was defined that is directed to the editors of BP descriptions.
The objective of the guidelines is to let the editors be aware of the expected quality of their
content. Such guidelines take inspiration from both the previously referred public guidelines
for defining PA documents, and from the Plain English Guide by Cutts (1996) (referred with
the acronym MC in the following). The guidelines were designed to be concise and clear,
and to act as an easily accessible checklist that the editors can read to verify that their BP
description has the appropriate degree of quality. For those guidelines that can be traced to
one of literature sources, the id of the source (EU, UK, USA, MC) is placed close to the
guideline, so that the interested reader can access the original manual, if she/he wishes to
have a major insight, or more specific examples.
The list of guidelines is reported in Table II, together with the defects of our domain
analysis that the guideline is expected to mitigate. The reader can verify that each defect is
mitigated by at least one guideline.
We derived a grouping of guidelines by clustering them iteratively as we discovered
them – based on similarity as perceived by the authors – and finally finding headlines
for clusters. Hence, guidelines below are partitioned into five groups, namely general
(i.e. guidelines that impact the whole procedure description), fields (i.e. guidelines
that specify the fields needed in a description – these guidelines are enforced by the
template described in Sect. 5.2), steps (i.e. guidelines associated to the partitioning
of the procedure into steps), sentences (i.e. guidelines associated to the writing style and
the clarity of the text), and warnings (i.e. guidelines associated to the specification of
exceptional situations).
In order to clarify the usage of part of these guidelines, we consider it helpful to illustrate
some of them with examples. In the following we have provided examples for those
guidelines concerning linguistic defects that are non-obvious; can be illustrated within a
single sentence; cannot be found in other manuals. For the examples, we adapted the text of
the real-world BP descriptions that we downloaded from the Web. Each row of Table III
refers to: the guideline (G-ID from Table II); the associated defect (D-ID from Table I);
an example of a defective sentence; and a suggestion for improving it. Additional examples
can be found in the manuals referenced in Table II.
54
BPMJ
24,1
D
ow
nl
oa
de
d 
by
 F
A
CH
H
O
CH
SC
H
U
LE
 N
O
RD
W
ES
TS
CH
W
EI
Z 
A
t 0
2:
03
 0
7 
M
ar
ch
 2
01
8 
(P
T)
D-ID Defect Source
Clarity of documents
1.a The document does not have an explicit argument/topic I2
1.b The document does not have an explicit motivation/function I4
1.c The document or the sections do not have a clear title EU
1.d The document is not partitioned into sections I1
1.e The document does not include practical examples I1
1.f The document does not explain the motivation of specific rules or instructions I4
1.g There is no glossary EU
1.h There is no reference to the software tools that shall be used I2
Clarity of sentences
2.a Sentences use a difficult language UK
2.b Sentences are too long UK
2.c Sentences include too many different concepts USA
2.d Rules/instructions are difficult to understand UK
2.e Rules/instructions are difficult to put into practice I4
2.f Relevant terms do not have a clear definition EU
2.g Sentences have an ambiguous structure I1
2.h Sentences use ambiguous terms I1
2.i Sentences use terms that are typical of other offices I1
2.j Sentences use too many synonyms I4
2.k Sentences contain grammatical errors I1
2.l Sentences contain juridical jargon UK
2.m Terms have specific meanings but they are used inappropriately I2
2.n Acronyms and abbreviations are not defined UK
Synthesis
3.a The document contains too many references to laws and regulations I1
3.b The document contains repetitions I2
3.c The document does not make clear what is important and what is not I1
3.d The document contains lists of steps that are too long UK
3.e The document is too long I3
3.f The document is too detailed I3
3.g The document contains obvious information EU
3.h The document refers irrelevant information I3
Internal coherence
4.a Relevant information is missing from the document I4
4.b The document does not make clear what are the institutions/offices involved I2
4.c The document does not explain who are the subjects involved I2
4.d The recipient of the document is unclear I2
4.e The document describes a procedure, but there is no explicit sequence of steps to be carried out USA
4.f The document describes a procedure with an explicit sequence of steps, but some steps are missing I4
4.g The document describes a procedure with an explicit sequence of steps, but the sequence is illogical USA
4.h The document does not explain what to do/who to contact if a problem occurs I1
4.i The document contains parts that contradict each other I2
4.j The document defines constraints that are too strict I1
4.k The document defines constraints that are illogical according to common sense EU
4.l The document leaves too much room for individual choices I4
4.m The structure of the document is inconsistent I2
External coherence
5.a The document does not include enough context information I1
5.b The document combines instructions with context information I4
5.c The document does not have an explicit category I2
(continued )
Table I.
List of defects from
the domain analysis
55
Quality of
business
process
descriptions
D
ow
nl
oa
de
d 
by
 F
A
CH
H
O
CH
SC
H
U
LE
 N
O
RD
W
ES
TS
CH
W
EI
Z 
A
t 0
2:
03
 0
7 
M
ar
ch
 2
01
8 
(P
T)
5.2 Template
According to the defects identified during the domain analysis, we defined a template for
editing BP descriptions, which can be used by editors of the PA to improve the clarity of
their content. The fields of the template, together with a short description, are reported
below. Moreover, for each field, we refer the defects that are expected to be mitigated by
that field:
 category [5.c]: the category of the document;
 headline [1.a, 1.c]: a short title describing the content;
 source documents [3.a]: identifiers of norms, regulations or any other document from
which the BP has been derived;
 reference documents [3.a]: identifiers of norms, regulations or any other document
that might have an impact on the current description;
 glossary [1.g, 2.n]: list of definitions that are useful to understand the BP;
 context [5.a, 5.b]: a brief overview of the information that might be useful for a reader
to understand the current BP;
 summary [3.d-3.f]: brief summary of the BP;
 motivation [1.b]: the higher-level objective of the BP;
 intended readership [4.d]: type of roles that should read this NL content;
 involved actors [4.b, 4.c]: actors (e.g. people, offices, authorities, etc.) that are involved
in the BP;
 input documents [3.a]: documents used as input for the current BP, if any;
 output documents [3.a]: documents produced by the current BP, if any;
 required tools [1.h]: software or hardware tools to be used to perform the BP, if any;
 instructions [4.e, 4.b]: actual description of the BP, expressed in the form of steps;
 examples/experiences [1.e]: list of real-world examples to practically describe the BP;
 what to do in case of failures [4.h]: suggestions of possible alternative choices to take
if something goes wrong while performing the current BP;
 contacts of involved offices [4.b]: name, phone number and e-mail of the offices
involved in the BP;
 contacts of experts [4.h]: name, phone number and e-mail of the BP experts to contact
to ask for clarifications; and
 FAQ [4.h]: list of frequently asked questions.
D-ID Defect Source
5.d Documents in the same category do not have a uniform structure I2
5.e There is inconsistency between the category of the document and its contents I2
5.f The document defines rules or procedures that overlap with other documents I2
5.g The document contradicts other documents I2
5.h The document contains copy-paste errors EU
5.i The document refers to other documents without justifying the reference I4
5.j The document does not mention other important documents I1Table I.
56
BPMJ
24,1
D
ow
nl
oa
de
d 
by
 F
A
CH
H
O
CH
SC
H
U
LE
 N
O
RD
W
ES
TS
CH
W
EI
Z 
A
t 0
2:
03
 0
7 
M
ar
ch
 2
01
8 
(P
T)
G-ID Guideline Defect
General
1.1 Be clear, concise and coherent (EU) 3.e-h, 4.i
1.2 Divide the procedure into steps (US) 4.e
1.3 Motivate the procedure, the steps and the references 1.b, 1.f, 4.k, 5.i
1.4 Leave the appropriate space for individual choices 4.j, 4.l
1.5 Do not describe obvious/common sense issues (EU) 3.g
1.6 Put it into practice what you wrote to check its applicability 2.e, 4.a, 4.e
1.7 Provide examples 1.e
1.8 Your procedure shall not contradict/overlap with other procedures 5.f-g
1.9 Take care that the structure of the document follows the structure of other
documents in the same category 5.d
1.10 Take care that the content of the document is consistent with its category 5.e
Fields
2.1 Provide a glossary (EU) 1.g, 2.f
2.2 Define an overview of the procedure 3d-f
2.3 Specify the topic and the category of the procedure 1.a, 5.c
2.4 Specify the scope/context of the procedure 5.a
2.5 Specify the intended reader of the procedure 4.d
2.6 Specify the subjects involved in the procedure 4.b-c
2.7 Partition the content into sections 1.d
2.8 Specify the tools needed to perform the procedure (web link, documents, software) 1.h
2.9 Reference only those documents that are strictly relevant 3.a, 5.j
2.10 Reference other relevant documents instead of repeating their content 5.f
Steps
3.1 Divide a procedure in logically linked steps (USA) 4.e, 4.g
3.2 Separate the steps with new lines 4.e
3.3 Define a label for each step 4.e
3.4 Use bullet points or numbered lists to identify the steps (USA) 4.g
3.5 If the chronological order of the steps is important, use a numbered list (USA) 4.g
3.6 If the order of the step is not important (steps can be performed in parallel), use
bullet points (USA) 4.g
3.7 Use action verbs in steps (Do, Make, Fill-out, etc.) (MC) 2.d-e
3.8 Use the imperative action verb at the beginning of each step (MC) 2.d-e
3.9 If conditions apply to the action, include them before the action verb (MC) 2.d-e
3.10 Do not mix instructions in steps with contextual information 5.b
3.11 Use 7-10 steps maximum for each procedure (UK) 3.d
3.12 If more than 7-10 steps are needed, partition the procedure into sub-tasks (UK) 3.d
3.13 Give clear and verbose headings (EU) 1.c
3.14 Use a consistent structure for all parts of the procedure 4.m
Sentences
4.1 Clarify acronyms and abbreviations (UK) 2.n
4.2 Highlight keywords and relevant content 3.c
4.3 Delete redundancies 3.b
4.4 Avoid grammatical errors 2.k
4.5 Avoid linguistic ambiguities in words and sentences 2.g-h
4.6 Use connectives (hence, therefore, etc.) between sentences (USA) 1.f
4.7 Use short sentences (max 25 words) (UK) 2.d, 2.b
4.8 Use short paragraphs (max 5 sentences) (UK) 1.d, 3.f
4.9 Cover only one topic per sentence / paragraph (USA) 2.c
4.10 Avoid double negations (USA) 2.a, 2.d
4.11 Keep subject, verb and object close together (USA) 2.a, 2.d
4.12 Use the word must for obligations (USA) 2.d
(continued )
Table II.
List of guidelines
for contributors of
the BP descriptions
57
Quality of
business
process
descriptions
D
ow
nl
oa
de
d 
by
 F
A
CH
H
O
CH
SC
H
U
LE
 N
O
RD
W
ES
TS
CH
W
EI
Z 
A
t 0
2:
03
 0
7 
M
ar
ch
 2
01
8 
(P
T)
6. Research challenges
The presented guidelines and the template are designed to address the problems of BP
descriptions, but, as any recommendation, rely on the good will of the civil servants who
edit the descriptions. Hence, to guarantee that a textual document has the required degree
of quality and does not include defects, other computer-aided approaches have to be put
into place to support the validation of the descriptions. We have identified four
macro-areas of research in which computer scientists can provide their contribution
towards a higher degree of quality of BP descriptions in the PA, namely readability
(Section 6.1), ambiguity (Section 6.2), relevance and text summarisation (Section 6.3),
modelling and consistency (Section 6.4).
6.1 Readability
According to our study, we have seen that, among the defects, appears the difficulty that the
civil servants encounter in reading their BP descriptions. Defects such as juridical jargon (2.j),
excessive length of the sentences (2.b) and other defects listed in the group named “clarity of
sentences” can be summarised as defects of language clarity. In general, this problem is
addressed by the computational linguistics field that goes under the name of readability.
Approaches to assess text readability focus mainly on the stylistic dimension,
i.e. assuring that a text can be easily read in terms of vocabulary, sentence structure and
other expressional elements by a certain group of readers (Dale and Chall, 1949). Early work
on automatic readability assessment was mainly focussed on defining formulas that could
associate a degree of readability to entire documents. In general, such readability formulas
assume that words with more syllables, and sentences with more words are less readable
(Kincaid et al., 1975; Coleman and Liau, 1975). Web-based implementations of these formulas
are also largely available online[10].
More recent studies on readability have started to take into account the actual content and
structure of sentences, based on the observation that a shorter term or a shorter sentence are
not necessarily more readable than longer ones (Dell’Orletta et al., 2014). In particular,
these works address the degree of difficulty of the vocabulary (Chall and Dale, 1995), and the
complexity of the syntax adopted in the documents (Petersen and Ostendorf, 2009; Aluisio
et al., 2010; Feng et al., 2010; Nenkova et al., 2010; Ma et al., 2012; Tanaka-Ishii et al., 2010).
All the cited works perform readability assessment at the document level – i.e. they provide a
G-ID Guideline Defect
4.13 Use verbs instead of nouns (evaluate instead of carry out an evaluation of ) (EU) 2.a, 2.d
4.14 Do not use synonyms for important terms 2.j
4.15 Do not use passive voice and name the subject who performs the action (EU) 4.c
4.16 Use adverbs only rarely 4.l
4.17 Avoid inconsistent use of terminology 2.m
4.18 Avoid inconsistent/contradictory content 4.i
4.19 Adapt the terminology to the target audience (EU) 2.i
4.20 When recycling text (copy/paste), make sure to properly adapt it (EU) 5.h
4.21 Do not use difficult terms (UK) 2.a, 2.d
4.22 Do not use juridical jargon (UK) 2.l
Warnings
5.1 Define warnings at the beginning, or before the step causing the warning (EU) 4.h
5.2 Tell the reader what to do if he/she makes a mistake 4.h
5.3 Include questions that you imagine the reader might have, and answer them (EU) 4.h
5.4 Specify people to contact in case of problems with the understanding of the procedure 4.h, 2.d
5.5 Specify people to contact in case of problems with the execution of the procedure 4.h, 2.eTable II.
58
BPMJ
24,1
D
ow
nl
oa
de
d 
by
 F
A
CH
H
O
CH
SC
H
U
LE
 N
O
RD
W
ES
TS
CH
W
EI
Z 
A
t 0
2:
03
 0
7 
M
ar
ch
 2
01
8 
(P
T)
readability measure, or rank for the entire document – and not at the sentence level.
These approaches help little in our context, where guidance is needed regarding the
improvement of specific parts of a text. Hence, the challenge here for computational linguists
resides in the need to establish more fine-grained techniques that are able to quantify
the readability of a single sentence, and in identifying potential readability defects. In this
sense, recent research in text simplification (François and Bernhard, 2014) can provide useful
outcomes that can be in principle employed also in the PA context. Nevertheless, one should
keep in mind that one cannot completely discard juridical jargon, which in principle hampers
readability. Indeed, the usage of juridical terms historically emerged for a need of precision,
which cannot be neglected. On the other hand, juridical jargon also implies complex sentence
structures, which can in principle be replaced with more readable sentences. Hence, our vision
is that research on readability of PA documents shall primarily focus on detecting readability
defects at the level of sentence structure rather than at the level of terminology.
G-ID/D-ID Example defect Suggested improvement
3.7/2.e The supervisor needs to be aware of any
extended periods of time that an employee will
be unavailable
The employee must notify the supervisor of any
extended period of time during which the
employee will be unavailable
3.9/2.e Projects with a sufficient average score will be
recommended for approval or recommended for
approval under conditions
If the average score of a project exceeds Y, it
will be recommended for approval. If it is
between X and Y, it will be recommended for
approval under conditions
3.10/5.b Inspectors may find it useful to review some of
their interventions, perhaps using the five
elements listed here as a guide, to consider
whether alternative approaches could have
been more effective
Inspectors should review their interventions
and consider the effectiveness of alternative
approaches. The five elements below may serve
as guidance for this: […]
3.13/1.c Application process How to apply to become a CDC
4.5/2.h The field office will forward the application to
the appropriate official for a final decision.
(vagueness)
The field office will forward the application to the
official who is responsible for the given area of
operations. The official will take the final decision
4.5/2.g The employee met the council and the head of
office and the secretary assessed his presence.
(syntactic, coordination)
The employee met the council. The head of
office and the secretary assessed his presence
4.5/2.g The delegate assesses the presence of the
candidate, and he provides his signature.
(syntactic, anaphoric)
The delegate assesses the presence of the
candidate and provides his signature
4.9/2.c Applications […] must be submitted or
authorised by the co-ordinating investigator (CI)
of the study in question who may delegate
responsibility for completing the application
form and uploading study documents to other
members of the study team
Applications […] must be submitted or
authorised by the co-ordinating investigator (CI)
of the study in question. The CI may delegate
responsibility for completing the application
form and uploading study documents to other
members of the study team
4.10/2.a If the committee does not disagree about
candidate selection […]
If the committee agrees on a candidate […]
4.11/2.d LEAP Academy will, according to the rules set
forth in N.J.A.C. 6:3-9, submit to the
Commissioner of Education an enrolment report
for the forthcoming year by June 1
LEAP Academy will submit an enrolment report
to the Commissioner of Education for the
forthcoming year by June 1. They need to do this
according to the rules set forth in N.J.A.C. 6:3-9
4.15/4.c The procedure shall be carried out before the
end of March 2015
The certification authority shall carry out the
procedure before the end of March 2015
4.21/2.a Review of decisions eligible for appeal Review of decisions allowed for appeal
4.22/2.1 To disagree with the dissolution petition fill in
the acknowledgement of service form
To disagree with the divorce petition fill in the
acknowledgement of service form
Table III.
Illustration of
relevant guidelines
with examples
59
Quality of
business
process
descriptions
D
ow
nl
oa
de
d 
by
 F
A
CH
H
O
CH
SC
H
U
LE
 N
O
RD
W
ES
TS
CH
W
EI
Z 
A
t 0
2:
03
 0
7 
M
ar
ch
 2
01
8 
(P
T)
6.2 Ambiguity
Still in the group that we named “clarity of sentences”, we have defects that refer to
ambiguity (2.g, 2.h). Ambiguity of terms and sentences occurs whenever the meaning
intended by the information producer (i.e. the writer) differs from the meaning understood
by the information consumer (i.e. the reader) (Ferrari et al., 2015). Ambiguity of terms is
associated to the so-called word-sense disambiguation (WSD) task (Navigli, 2009).
Techniques for WSD aim at identifying the intended meaning of a polysemous term – i.e.
a term with multiple meaning such as “bass” (which can indicate “bass guitar” or a kind of
fish) – depending on its linguistic context. Several approaches exist that address this
problem, which use unsupervised (Agirre and Edmonds, 2007; Véronis, 2004), supervised
(Lee and Ng, 2002) and knowledge-based approaches (Banerjee and Pedersen, 2003; Navigli
and Velardi, 2005). However, such techniques are mainly aimed to support information
retrieval (IR) and machine translation, and are not oriented to detect ambiguity as a defect.
Indeed, given a term, currently available tools are oriented to produce a potential sense
among the possible ones, and do not associate a degree of ambiguity to the term.
Instead, ambiguity as a quality defect has been largely studied in the field of
requirements engineering, in which requirements – i.e. statements indicating properties and
behaviour of a software – need to be understood by different stakeholders involved in the
development of a system – e.g., customers, developers and testers – and should be free of
ambiguity to avoid misunderstandings about desired system properties. Some studies
address the identification of typical ambiguous terms and constructions (Berry et al., 2003;
Gnesi et al., 2005; Gleich et al., 2010). Other work addresses ambiguity by translating
requirements into formal languages or models (Cimatti et al., 2011; Ambriola and Gervasi,
2006; Kof, 2010). Finally, some work focusses on the usage of natural language
understanding methodologies (Mich and Garigliano, 2000; Kiyavitskaya et al., 2007).
The most mature works in this field, which can be employed also for PA documents,
concern: the identification of vague expressions (Gleich et al., 2010; Gnesi et al., 2005), such as
“as soon as possible”, “appropriate” – see example 4.5/2.h in Table III; the identification
of syntactic ambiguities, such as coordination (Chantree et al., 2006) and anaphoric
(Yang et al., 2011) ones – see examples 4.5/2.g in Table III.
Besides requirements engineering, ambiguity has also been studied for the domain of legal
texts, where different interpretations of, e.g., contracts or statutes, lead to problems in jurisdiction
(Solan, 2004; Solan et al., 2008) and in interpretation of regulations (Massey et al., 2014).
Especially syntactic ambiguity has been addressed as a problem in technical texts
intended for (cheap) translation where human translators have to spot such ambiguities
when post-editing machine translations; solutions are based e.g., on so-called “syntactical
cues” as discussed by Kohl (1999).
Overall, we believe that research on ambiguity in PA documents can leverage the
literature concerning vagueness and syntactic ambiguity in requirements. For lexical
ambiguity – i.e. when a term has more than one vocabulary meaning – inspiration can be
taken from works on Word Sense Disambiguation, but such works shall be tailored for
defect identification.
6.3 Relevance and text summarisation
Most of the defects listed in the group named “synthesis”, are due to the presence of some form
of redundancy, e.g., due to the excessive number of external references (3.a), or to the absence
of clarification of what is relevant and what is not (3.c). Addressing these problems with
automated means is not straightforward. Indeed, all these defect deal with the vague concept
of relevance. Relevance is studied in the field of IR, and several notions
are introduced in the literature (Borlund, 2003; Huang and Soergel, 2013), from topic relevance
(i.e. relevance of an information item with respect to a certain topic), to evidential relevance
60
BPMJ
24,1
D
ow
nl
oa
de
d 
by
 F
A
CH
H
O
CH
SC
H
U
LE
 N
O
RD
W
ES
TS
CH
W
EI
Z 
A
t 0
2:
03
 0
7 
M
ar
ch
 2
01
8 
(P
T)
(i.e. relevance of an information item to reason about a certain fact), to situational relevance
(i.e. relevance of an information item for performing a specific task). In the case of BP
descriptions, research should mainly focus on situational relevance. Indeed, these descriptions
should be used by civil servants to help them performing specific tasks, and it is therefore
important to devise strategies that allow emphasising those parts of the text that provide
practically usable information, i.e. information that can be used to perform a task.
To address this goal, research should look at works on automatic text summarisation
(Nenkova and McKeown, 2012; Lloret and Palomar, 2012). Text summarisation techniques
are distinguished in extractive and abstractive (Lloret and Palomar, 2012). Extractive
approaches score the sentences of a document based on their computed relevance, and
produce a summary that includes a subset of relevant sentences. Abstractive approaches
attempt to manipulate the original content to produce a summary, with a particular focus on
the coherence of the produced summary. In case of BP descriptions, extractive approaches
such as those listed by Gupta and Lehal (2010), can be useful to identify which are the
relevant sentences in a BP description, and suggest the editor which ones should
be highlighted. Moreover, these approaches can be also employed to suggest which are the
redundant sentences that could be discarded. Abstractive approaches could be employed to
produce actual summaries, to be placed at the beginning of the BP descriptions.
All these approaches can be adapted to the domain of BP descriptions. However, as noted
by Ferreira et al. (2013), the effectiveness of automatic summarisation highly depends on the
subject of the text that is used. Hence, appropriate experiments on real-world BP description
shall be carried out to assess the applicability of these methods.
6.4 Modelling and consistency
Part of the problems listed in the groups of “internal coherence” and “external coherence”
are addressed by means of our template (e.g. 4.b-d and 4.h, referring to specific actors to be
mentioned, or 5.c, referring to the document category). Nevertheless, some of the relevant
problems concern the logical consistency (e.g. 4.f-g, 4.i) of the BP description, and
these problems cannot be automatically addressed only looking at the text. To address these
problems, more rigorous representations of the BPs are required. In this sense, the BPMN
notation (OMG, 2011) is a natural candidate to model BPs in the PA.
In the PA domain, studies were performed, e.g., in the Spanish PA by Torres et al. (2010),
and in the Swiss PA by Schumann et al. (2014). Moreover, BPMN was used also to model
heath-care processes – which can be regarded as special cases of PA processes – by Müller
and Rogge-Solti (2011) and by Ruiz et al. (2012). These works generally aim at showing the
effectiveness of a rigorous modelling approach in a human-intensive regulated environment,
with the goal of easing decision making. Other works make a step forward, and focus on the
analysis of the BPMN models of the PA (Falcioni et al., 2012; Bruni et al., 2011).
Overall, these works can in principle address inconsistency problems raised in our
domain analysis (e.g. 4.f-g, 4.i). However, as noted by Ottensooser et al. (2012), and Nawrocki
et al. (2006), since BP descriptions in the PA are going to be complemented with textual ones,
problems of inconsistency, although discarded in the models, might still remain in the text.
Therefore, we argue that technologies shall be devised to identify the consistency between
the textual procedure description, and the corresponding model. In this sense, approaches
exist that are able to generate textual descriptions from BPMN models. In particular,
Leopold et al. (2014) presents a tool-supported approach for natural language generation of
BP descriptions from BPMN models. In this way, the consistency of the NL description is
guaranteed by construction, assuming that the original model is consistent. On the other
hand, research shall also look at other lightweight approaches to ensure model-to-text
consistency, since, in some cases, a manually edited BP description might cover aspects that
are not covered in the model. In this case, one wants to check that the degree of alignment
61
Quality of
business
process
descriptions
D
ow
nl
oa
de
d 
by
 F
A
CH
H
O
CH
SC
H
U
LE
 N
O
RD
W
ES
TS
CH
W
EI
Z 
A
t 0
2:
03
 0
7 
M
ar
ch
 2
01
8 
(P
T)
between BP description and model, e.g., to see if relevant entities mentioned in the model are
also mentioned in the text, and vice-versa. To address this goal, research should look at
terminology extraction techniques (Medelyan et al., 2013). Extracting the relevant terms of
the BP descriptions and comparing them with the labels in the BPMN models can provide a
lightweight strategy to check existing relations between text and model.
Of course, not all the problems can be addressed through modelling. For example,
steps that are illogical according to common sense (4.k), or constraints that are too strict
or too weak (4.j, 4.l) cannot be automatically identified through model analysis. However,
we argue that the sole fact of having a model, which implies having a rigorous and
concise representation of the BP, can provide a clearer perspective on which are the illogical
parts of a BP.
7. Conclusions and future work
Effective communication of BP knowledge is a primary need for any structured
organisation. This is particularly true for PAs, in which poor communication of BP
knowledge to civil servants may hamper the efficiency of the bureaucracy, and lead to
citizens’ distrust towards the means of their state.
In this paper, we provide an insight on the current problems in PA documents, in which
part of the BP knowledge is currently conveyed, and on solutions to improve the quality of
such documents. To identify the problems of current PA textual documents, we present a
thorough domain analysis, comprising a literature review and interviews that involve
PA stakeholders. To define the solutions, we provide a PA-specific template for editing
BP descriptions, which is oriented to reduce a large part of the problems identified, and a set
of editing guidelines. Moreover, we review the relevant literature in different fields, namely
readability, ambiguity, relevance and modelling, and we highlight how the contributions
of the current literature can be applied to automatically address the defects that our research
revealed. Our near future commitment is the development of specific techniques to address
readability and ambiguity problems in BP descriptions. Moreover, we plan to employ
modelling techniques based on BPMN, and we plan to devise strategies to verify
the alignment between textual BPs and BP models. We foresee that validation strategies
based on this dual representation of a BP (natural language and model) can mitigate
coherence and synthesis defects. From a practical standpoint, we foresee that a
higher degree of automation in text quality assurance will speed-up the BPs of PAs,
by reducing communication problems, and by establishing a clearer way of communication
between BP stakeholders.
Notes
1. www.plainenglishfoundation.com
2. www.plainenglish.co.uk
3. www.plainlanguagenetwork.org
4. www.hemingwayapp.com
5. http://goo.gl/Wx1Tef
6. www.plainlanguage.gov
7. www.isti.cnr.it
8. www.interreg2seas.eu/en/program
9. www.impresainungiorno.gov.it/psc-italy
10. https://readability-score.com
62
BPMJ
24,1
D
ow
nl
oa
de
d 
by
 F
A
CH
H
O
CH
SC
H
U
LE
 N
O
RD
W
ES
TS
CH
W
EI
Z 
A
t 0
2:
03
 0
7 
M
ar
ch
 2
01
8 
(P
T)
References
Agirre, E. and Edmonds, P.G. (2007), Word Sense Disambiguation: Algorithms and Applications, Vol. 33,
Springer Science & Business Media, New York, NY.
Aluisio, S., Specia, L., Gasperin, C. and Scarton, C. (2010), “Readability assessment for text
simplification”, Proceedings of the NAACL HLT 2010 5th Workshop on Innovative Use
of NLP for Building Educational Applications. Association for Computational Linguistics,
pp. 1-9.
Ambriola, V. and Gervasi, V. (2006), “On the systematic analysis of natural language requirements
with Circe”, Automated Software Engineering, Vol. 13 No. 1, pp. 107-167.
Banerjee, S. and Pedersen, T. (2003), “Extended gloss overlaps as a measure of semantic relatedness”,
Proceedings of the 18th International Joint Conference on Artificial Intelligence (IJCAI), Vol. 3,
pp. 805-810.
Batini, C. and Scannapieco, M. (2016), “Information quality dimensions for maps and texts”, in Batini, C.
and Scannapieco, M. (Eds), Data and Information Quality, Springer International Publishing
Switzerland, Cham, pp. 53-86.
Berry, D., Gacitua, R., Sawyer, P. and Tjong, S.F. (2012), “The case for dumb requirements engineering
tools”, International Working Conference on Requirements Engineering: Foundation for Software
Quality, Springer, pp. 211-217.
Berry, D.M., Kamsties, E. and Krieger, M.M. (2003), “From contract drafting to software specification:
linguistic sources of ambiguity”.
Borlund, P. (2003), “The concept of relevance in IR”, Journal of the American Society for Information
Science and Technology, Vol. 54 No. 10, pp. 913-925.
Bruni, R., Corradini, A., Ferrari, G., Flagella, T., Guanciale, R. and Spagnolo, G. (2011), “Applying process
analysis to the italian egovernment enterprise architecture”, in Carbone, M. and Petit, J.-M. (Eds),
Web Services and Formal Methods, Springer-Verlag, Berlin and Heidelberg, pp. 111-127.
Buican, I., Hriscu, V. and Amador, M. (1993), “Using International english to prepare technical text for
translation. in professional communication conference, 1993”, IPCC 93 Proceedings. The New
Face of Technical Communication: People, Processes, Products, IEEE, pp. 33-35.
Chall, J.S. and Dale, E. (1995), Readability Revisited: The New Dale-Chall Readability Formula, Brookline
Books, Northampton, MA.
Chantree, F., Nuseibeh, B., Roeck, A.N.D. and Willis, A. (2006), “Identifying nocuous ambiguities in
natural language requirements”, Proceeding of RE’06, pp. 56-65.
Cimatti, A., Roveri, M., Susi, A. and Tonetta, S. (2011), “Formalizing requirements with object models
and temporal constraints”, Software and Systems Modeling, Vol. 10 No. 2, pp. 147-160.
Coleman, M. and Liau, T.L. (1975), “A computer readability formula designed for machine scoring”,
Journal of Applied Psychology, Vol. 60 No. 2, pp. 283-284.
Cutts, M. (1996), The Plain English Guide, Oxford University Press, Oxford.
Dale, E. and Chall, J.S. (1949), “The concept of readability”, Elementary English, Vol. 26 No. 1, pp. 19-26.
Dell’Orletta, F., Montemagni, S. and Venturi, G. (2014), “Assessing document and sentence readability
in less resourced languages and across textual genres”, International Journal of Applied
Linguistics, Vol. 165 No. 2, pp. 163-193.
Di Ciccio, C., Marrella, A. and Russo, A. (2015), “Knowledge-intensive processes: characteristics,
requirements and analysis of contemporary approaches”, Journal on Data Semantics, Vol. 4
No. 1, pp. 29-57.
Di Cicco-Bloom, B. and Crabtree, B.F. (2006), “The qualitative research interview”, Medical Education,
Vol. 40 No. 4, pp. 314-321.
European Commission (2011), English Style Guide: a handbook for authors and translators in the
European Commission, 7th ed., available at: http://ec.europa.eu/translation/english/guidelines/
documents/ styleguide_english_dgt_en.pdf (accessed 18 September 2015).
63
Quality of
business
process
descriptions
D
ow
nl
oa
de
d 
by
 F
A
CH
H
O
CH
SC
H
U
LE
 N
O
RD
W
ES
TS
CH
W
EI
Z 
A
t 0
2:
03
 0
7 
M
ar
ch
 2
01
8 
(P
T)
European Commission (2012), “How to write clearly”, doi: 10.2782/54318.
European Commission (2014), “Claire’s clear writing tips”, available at: http://ec.europa. eu/translation/
documents/clear_writing_tips_en.pdf (accessed 18 September 2015).
Falcioni, D., Polini, A., Polzonetti, A. and Re, B. (2012), “Improving PA business processes through
modeling, analysis, and reengineering”, in Scholl, H.J., Janssen, M., Wimmer, M.A., Moe, C.E. and
Flak, L.S. (Eds), Electronic Government, Springer, Berlin, pp. 78-89.
Feng, L., Jansche, M., Huenerfauth, M. and Elhadad, N. (2010), “A comparison of features for automatic
readability assessment”, Proceedings of the 23rd International Conference on Computational
Linguistics: Posters. Association for Computational Linguistics, pp. 276-284.
Ferrari, A., Spoletini, P. and Gnesi, S. (2015), “Ambiguity as a resource to disclose tacit knowledge”,
Proceedings of the 23rd International Requirements Engineering Conference on (RE) IEEE,
pp. 26-35.
Ferreira, R., de Souza Cabral, L., Lins, R.D., e Silva, G.P., Freitas, F., Cavalcanti, G.D., Lima, R., Simske, S.J.
and Favaro, L. (2013), “Assessing sentence scoring techniques for extractive text summarization”,
Expert Systems with Applications, Vol. 40 No. 14, pp. 5755-5764.
François, T. and Bernhard, D. (2014), “Recent advances in automatic readability assessment and text
simplification”, ITL: International Journal of Applied Linguistics, Vol. 165 No. 2.
Gleich, B., Creighton, O. and Kof, L. (2010), “Ambiguity detection: towards a tool explaining ambiguity
sources”, Proceeding of REFSQ’10. Springer Volume 6182 of LNCS, pp. 218-232.
Gnesi, S., Lami, G. and Trentanni, G. (2005), “An automatic tool for the analysis of natural language
requirements”, Computer Systems Science & Engineering, Vol. 20 No. 1, 13pp.
Gupta, V. and Lehal, G.S. (2010), “A survey of text summarization extractive techniques”, Journal of
Emerging Technologies in Web Intelligence, Vol. 2 No. 3, pp. 258-268.
Huang, X. and Soergel, D. (2013), “Relevance: an improved framework for explicating the notion”,
Journal of the American Society for Information Science and Technology, Vol. 64 No. 1,
pp. 18-35.
Kincaid, J.P., Fishburne, R.P. Jr, Rogers, R.L. and Chissom, B.S. (1975), “Derivation of new readability
formulas (automated readability index, fog count and flesch reading ease formula) for navy
enlisted personnel”, technical report, DTIC Document, Fort Belvoir, VA.
Kiyavitskaya, N., Zeni, N., Mich, L. and Berry, D.M. (2007), “Requirements for tools for ambiguity
identification and measurement in natural language requirements specifications”, Proceedings of
WER’07, pp. 197-206.
Kof, L. (2010), “From requirements documents to system models: a tool for interactive semi-automatic
translation”, Proceedings of the 18th International Requirements Engineering Conference (RE)
IEEE, Sydney, 27 September-1 October, pp. 391-391.
Kohl, J.R. (1999), “Improving translatability and readability with syntactic cues”, Technical
Communication, Vol. 46 No. 2, pp. 149-166.
Lee, Y.K. and Ng, H.T. (2002), “An empirical evaluation of knowledge sources and learning
algorithms for word sense disambiguation”, Proceedings of the ACL-02 Conference on Empirical
Methods in Natural Language Processing – Vol. 10. Association for Computational Linguistics,
pp. 41-48.
Leopold, H., Mendling, J. and Polyvyanyy, A. (2014), “Supporting process model validation through
natural language generation”, Software Engineering, IEEE Transactions on, Vol. 40 No. 8,
pp. 818-840.
Lloret, E. and Palomar, M. (2012), “Text summarisation in progress: a literature review”, Artificial
Intelligence Review, Vol. 37 No. 1, pp. 1-41.
Ma, Y., Fosler-Lussier, E. and Lofthus, R. (2012), “Ranking-based readability assessment for early
primary children’s literature”, Proceedings of the 2012 Conference of the North American
Chapter of the Association for Computational Linguistics: Human Language Technologies.
Association for Computational Linguistics, pp. 548-552.
64
BPMJ
24,1
D
ow
nl
oa
de
d 
by
 F
A
CH
H
O
CH
SC
H
U
LE
 N
O
RD
W
ES
TS
CH
W
EI
Z 
A
t 0
2:
03
 0
7 
M
ar
ch
 2
01
8 
(P
T)
Massey, A.K., Rutledge, R.L., Antón, A.I. and Swire, P.P. (2014), “Identifying and classifying ambiguity
for regulatory requirements”, Proceeedings of the 22nd International Requirements Engineering
Conference (RE) IEEE, pp. 83-92.
Medelyan, O., Witten, I.H., Divoli, A. and Broekstra, J. (2013), “Automatic construction of lexicons,
taxonomies, ontologies, and other knowledge structures”, Wiley Interdisciplinary Reviews: Data
Mining and Knowledge Discovery, Vol. 3 No. 4, pp. 257-279.
Mich, L. and Garigliano, R. (2000), “Ambiguity measures in requirements engineering”, Proceedings of
ICS’00, 16th IFIP WCC, pp. 39-48.
Mu ̈ller, R. and Rogge-Solti, A. (2011), “Bpmn for healthcare processes”, Proceedings of the 3rd Central-
European Workshop on Services and their Composition (ZEUS 2011), Karlsruhe.
Navigli, R. (2009), “Word sense disambiguation: a survey”,ACMComputing Surveys (CSUR), Vol. 41 No. 2,
Article No. 10.
Navigli, R. and Velardi, P. (2005), “Structural semantic interconnections: a knowledge-based approach to
word sense disambiguation”, Pattern Analysis and Machine Intelligence, IEEE Transactions on,
Vol. 27 No. 7, pp. 1075-1086.
Nawrocki, J.R., Nedza, T., Ochodek, M. and Olek, L. (2006), “Describing business processes with use
cases”, 9th International Conference on Business Information Systems (BIS), pp. 13-27.
Nenkova, A., Chae, J., Louis, A. and Pitler, E. (2010), “Structural features for predicting the linguistic
quality of text”, in Krahme, E. and Theune, M. (Eds), Empirical Methods in Natural Language
Generation, Springer-Verlag, Berlin and Heidelberg, pp. 222-241.
Nenkova, A. andMcKeown, K. (2012), “A survey of text summarization techniques”, in Aggarwal, C.C. and
Zhai, C.X. (Eds), Mining Text Data, Springer Science and Business Media, LLC, New York, NY,
pp. 43-76.
Object Management Group (OMG) (2011), “Business Process Model and Notation (BPMN) version
2.0”, technical report, Object Management Group, Needham, MA, available at: www.omg.org/
spec/BPMN/2.0/ (accessed 19 December 2017).
Ottensooser, A., Fekete, A., Reijers, H.A., Mendling, J. and Menictas, C. (2012), “Making sense of
business process descriptions: an experimental comparison of graphical and textual notations”,
Journal of Systems and Software, Vol. 85 No. 3, pp. 596-606.
Petersen, S.E. and Ostendorf, M. (2009), “A machine learning approach to reading level assessment”,
Computer Speech & Language, Vol. 23 No. 1, pp. 89-106.
Redish, J.G. (2012), Letting Go of the Words: Writing Web Content that Works, Elsevier, Amsterdam.
Ruiz, F., Garcia, F., Calahorra, L., Llorente, C., Gonçalves, L., Daniel, C. and Blobel, B. (2012), “Business
process modeling in healthcare”, Stud Health Technol Inform, Vol. 179, pp. 75-87.
Sánchez-González, L., Ruiz, F., García, F. and Cardoso, J. (2011), “Towards thresholds of control flow
complexity measures for BPMN models”, Proceedings of the 2011 ACM Symposium on Applied
Computing SAC ’11. ACM, New York, NY, pp. 1445-1450, available at: http://doi.acm.org/10.
1145/1982185.1982496
Schumann, R., Delafontaine, S., Taramarcaz, C. and Evéquoz, F. (2014), “Effective Business process
documentation in federal structures”, No. 44, Jahrestagung der Gesellschaft für Informatik,
pp. 1043-1057.
Solan, L.M. (2004), “Pernicious ambiguity in contracts and statutes”, ChicagoKent Law Review,
Vol. 79 No. 3, pp. 859-888.
Solan, L.M., Rosenblatt, T. and Osherson, D. (2008), “False consensus bias in contract interpretation”,
Columbia Law Review, Vol. 108, pp. 1268-1300.
Sommerville, I. (2007), “Models for responsibility assignment”, in Dewsbury, G. and Dobson, J. (Eds),
Responsibility and Dependable Systems, Springer, Berlin, pp. 165-186.
Tanaka-Ishii, K., Tezuka, S. and Terada, H. (2010), “Sorting texts by readability”, Computational
Linguistics, Vol. 36 No. 2, pp. 203-227.
65
Quality of
business
process
descriptions
D
ow
nl
oa
de
d 
by
 F
A
CH
H
O
CH
SC
H
U
LE
 N
O
RD
W
ES
TS
CH
W
EI
Z 
A
t 0
2:
03
 0
7 
M
ar
ch
 2
01
8 
(P
T)
Torres, V., Giner, P., Bonet, B. and Pelechano, V. (2010), “Adapting BPMN to public administration”,
in Mendling, J., Weidlich, M. and Weske, M. (Eds), Business Process Modeling Notation,
Springer-Verlag, Berlin and Heidelberg, pp. 114-120.
Trudeau, C.R. (2012), “The public speaks: an empirical study of legal communication”, The Scribes
J. Leg. Writing, 14.
UK Government (2015), “Content design: planning, writing and managing content”, available at:
www.gov.uk/guidance/content-design/ writing-for-gov-uk (accessed 1 August 2015).
US Government (2011), “Federal plain language guidelines”, available at: https://plainlanguage.gov/
media/FederalPLGuidelines.pdf (accessed 19 December 2017).
Véronis, J. (2004), “Hyperlex: lexical cartography for information retrieval”, Computer Speech &
Language, Vol. 18 No. 3, pp. 223-252.
Williams, J. and Colomb, G. (1995), Style: Toward Clarity and Grace. Chicago guides to Writing, Editing,
and Publishing, University of Chicago Press, Chicago, IL.
Yang, H., Roeck, A.N.D., Gervasi, V., Willis, A. and Nuseibeh, B. (2011), “Analysing anaphoric ambiguity
in natural language requirements”, Requirements Engineering, Vol. 16 No. 3, pp. 163-189.
Corresponding author
Alessio Ferrari can be contacted at: alessio.ferrari@isti.cnr.it
For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com
66
BPMJ
24,1
D
ow
nl
oa
de
d 
by
 F
A
CH
H
O
CH
SC
H
U
LE
 N
O
RD
W
ES
TS
CH
W
EI
Z 
A
t 0
2:
03
 0
7 
M
ar
ch
 2
01
8 
(P
T)