Gérald Péoux is an assistant professor at the Université Paris Ouest Nanterre La Défense, where he gives lectures in computer science, and a researcher in the history of Science at IHMC (Institut d’Histoire Moderne et Contemporaine). Before joining the university, he had a position as an engineer in supercomputers in private American companies.
Jean-Roch Houllier, PMP, IPMA Level B certified, MGP, SSCBB, FFP,
Executive Coach Certified HEC Paris, is a graduate of HEC Business
School and SUPAERO. He is currently THALES
Jean-Roch is passionate about teaching and professional learning; he likes to share his experiences and encourage his students. He is a lecturer, teacher and Professional Thesis Director for HEC Business School in project management and the Academic Development Director of the PMI-France, developing contacts and partnerships with various schools around project management, making bridges between academic and professional universes.
This is the source
In this article, we discuss a way for visualizing interactions or solidarities in a past community. The method that we expound is derived from process-oriented views used in the industry. Therefore, this attempt to adopt practices from outside of the traditional humanities framework must be clearly justified in terms of epistemological position and in terms of choice among the numerous available tools. First, we try to define what the process term could mean in the humanities; then we explain our choice of a representation; we also carry out tests of our method on selected case studies to visualize and question scholarly works. Last, we offer the reader a prototype that must be improved to produce automatic visualization of particular situations.
A
Designing tools that help visualize scholarly works and offer various perspectives is undoubtedly one of the key features of the Digital Humanities. Indeed, as explained by members of the
we did not see beforeand
get a better sense of the overall shape and structure.
In this work, we discuss a
At first view there seems to be a tremendous distance between research on history and the search for optimal processes that would minimize the ratios of breakdowns and subsequently provide the high level of quality of an industrial production. Yet, our will to learn from the work of one another has convinced us that the notion of process was already present at various periods of time in history. Since in the industrial context people have a long-term experience on how to represent dynamic process-oriented situations from a start point to an end point, we argue that, by means of a proper translation, it is then possible to offer new perspectives on historical works, as long as they contain one or more process dimensions. Furthermore, we assert that these perspectives pertain to the Digital Humanities domain. Indeed, as Daniel Cohen says in his book
newtechnologies
First, we would like to justify that assumption by reviewing works carried out by historians and by showing that they could be visualized in terms of process. Then we shall present some process-oriented representations and we will explain the choice of one of them in particular. We will try to apply our methodology to two particular works, dealing for the first one with information management and for the second one with the industrial production of fabric. Last, dealing with technicalities, we will show a way to translate our representation in a technical fashion with the help of the last web standards and by presenting a client-side prototype based on a specialized
In this work, we define a process as a set of activities bringing about a result, performed with accuracy and full control by actors who may interact with each other as the links of the same chain. This chain of connected activities forms in a long or short period of time, consciously or not. All actors of such a process pursue a common objective which implies solidarities and sociability. Therefore, we argue that if we manage to disclose such processes from the past, we obtain information about the way communities were organized. Thus, our purpose in this article is to propose means and rigorous methodologies to visualize such processes in history. We chose to rest upon works of historians to illustrate our method, although we could also apply it to primary sources. Indeed, once we have identified relevant processes to highlight interactions between actors, we are ready to visualize them with the help of our tool and subsequently to question our sources on the basis on what was revealed by that new display.
First, let us exemplify our point with cases that were studied by historians and that are related to the production of knowledge.
Many works of historians deal with the production of knowledge at various periods
(Antiquity, Middle Ages, Renaissance, Enlightenment…) and in different areas
(Europe, the Islamic world, China…). In this section, we suggest that some
scholarly works could be represented in terms of process if we can agree to a
definition that would be both wide enough to model different situations and
sufficiently accurate to take into account the specificity of the humanities. In
a book published in 2000, Tiziano Dorandi analyzes various aspects of writing in
Greek and Roman Antiquity from the draft to the book edition
In an article entitled
tax registers, ecclesiastical payment ledgers and municipal account books.From these sources, the author identifies several groups of actors who might have participated in the elaboration of the work: parchment makers, papermakers, scribes, illuminators and bookbinders. For each of these groups, the author gives precise information about its internal organization and its interaction with its social and economic environment. For example, we learn that the wages paid to skilled and experienced scribes represented a great amount in the total cost of the production of a manuscript. We understand that the scribes were organized as a team led by a coordinator who might also be in charge of
contracting with a music annotator, illuminator, bookbinder.If we see this situation in terms of process, we can assert that a process is something which has the property to be connected. Indeed, the different groups involved in the making of the psalter must achieve their own task and forward the result to the next group. This is of course a sequential view of the chain of processes which does not forbid that certain parts of the manufacturing can be led in parallel (for instance, the making of the cover and the copying of the manuscript).
Other works of historians dealing with the production of knowledge in other cultural areas could also be quoted. As an example, let us mention the study of Johannes Kurz who proposed in 2007 an article about
These foregoing examples show that a number of human activities can be modelled by a set of processes and from this point of view, it allows us to emphasize the relationships between individuals in a historical, social and technical context. Hence, our problem is to find a way of representing such processes in a Digital Humanities framework. In this view, let us cite some process representations that have already been tested and experienced in DH.
Many Digital Humanities programs were initiated in recent years, of which some highlighted connections between actors, geographical areas or events and the underlying processes. For instance, the
Another worthwhile example is the
unjustly neglectedauthors. Jockers determines the frequencies of certain words, works on patterns and tabulates his results by genre, date, nationality or origin of the authors. This work is inscribed in the
distant readingprogram as defined by Franco Moretti in his
pressure of social selection. As in Darwinian evolution, that transformation seems to be the result of a long-term
In contrast, in this essay, we intend to concentrate on the processes that human communities set up for a precise objective and that dynamically account for an action being made. In this way, what we propose can complement the tools and methods that we have just mentioned above (such as tree diagrams or network mapping) insofar as we focus on the process itself in such a way that we can disclose the atomic dimensions of the situations we investigate. Thus, we consider interactions between elements that may be either individuals or material resources or even more abstract entities such as activities. The choice of such interactions enables us to account for complex situations (behind which we find individuals, organizations or external constraints) that we describe with an identical framework whatever the details that we obtain from historical works.
Therefore, we think that modern methods of process representation could contribute to highlight the interactions between these various elements and, as such, that they would give a relevant description of past collective enterprises like, for example, the production of knowledge.
In order to make a choice of a process representation for visualizing scholarly works, we should first consider what a process means in our contemporary world and especially in the industry where the need for formal and rigorous methods is crucial.
In a modern wording, a process may be defined as a series of linked activities
and resources (either material or human) which transform input elements into
output ones to construct an interpretative
framework
for accounting situations described in scholarly works.
This do[es] not represent the past, but
our ideas about the past
and, in our case, ideas derived from
scholarly works categories to delimit a
conception of the world
is already an activity of modeling
There are several ways of representing processes. All of them correspond to
different points of view. We can quote, for instance, those focused on heuristic
conceptions, usually called mind maps. They are very relevant to account for
decision processes but do not seem to correctly represent situations and
activities already well experienced as those we would like to illustrate
here.
The
Although UML focuses on software and system designs, it can be extended to other
purposes.
the capability of understanding the internal business procedures.
As a matter of fact, we preferred a methodology of representation derived from
the quality of process and we will try to justify our choice further. The
quality of process ensures that, for a certain probability or more precisely at
a certain error rate, which must be minimized, a process will produce the
expected output element. The philosophy of such a representation consists first
in observing what already exists and then, upon these observations, in
establishing a diagnosis of what must be improved. For this purpose, such
methods provide robust ways of building a framework or a model for apprehending
reality. With this process representation, we can position ourselves as
observers of experienced situations and not as designers of a system, which we
cannot pretend to be when scrutinizing the past. In the industry, minimizing an
error rate may be characterized as a
Let us now briefly describe the methodology. Companies implement process referential and maps, which define the interactions between processes, show and reveal the overall coherence. It is a tool of communication and therefore it represents the synoptic view of all business processes (sequence and interactions between processes).
Processes and roles relating to these are generally independent of the
organization and functions. The method of process representation that we
consider here is inscribed in an overall toolbox usually declined in five phases
of which the main and key objective is to converge to a
In the reality of industry, the project manager must ensure that basic questions
such as the project relationship with the overall objectives
, the
nature of the problem
and the goal
have been understood by all
stakeholders. Hence, it follows a progressive approach, including multiple
steps, each of them being focused on a specific problematic.
During the so-called first phase, the project manager starts by understanding the current process and its related characteristics (process boundaries, the way products and services are actually delivered, etc.). For that, several tools are useful to visualize and point to the process that needs improvement, to identify start and stop points of the process under consideration, to identify the customers of the process and to ensure a shared understanding of key elements of the process, including what it looks like and what it should deliver. These tools are the
high-level process mapsand the
detailed process maps.
Then, the high-level process maps
and detailed process maps
representations propose a focus on the process itself detailing its activities,
main flows, and other elements.
The SIPOC, as a first representation, greatly helps to understand the process context. This is very important, as chains of process in companies are quite usual, meaning that the outputs of the first become the main inputs of the next one, etc. To be brief, we can state that, in the SIPOC wording,
When focusing on the process itself, the high-level process maps
and the
detailed process maps
representations are useful. An example of a
process map that we will use and detail further in the present article, is the
flowchart. A process flowchart
includes several key elements which are the main activities and milestones of
the process; the flows between activities, e.g. main inputs and outputs; the
roles associated with the various activities; the different elements supporting
the run of a given activity.
Figure 2 is an example of the flowchart template used to represent the various processes in the
To summarize our choice, we state that we intend to apply a method used in the
industry to describe and to visualize particular situations reported in
scholarly works. This method consists in coupling a generic and macroscopic
representation (SIPOC) with a more detailed process map (namely a flowchart)
specifically developed and thought of by the Thales Company.
Indeed, in their work on Slavery in American Society before the Civil War
conducted at the University of Virginia, W. G. Thomas and E. L. Ayers argue that
one of the activities of a Digital Humanities project is the translation of professional scholarship-evidence into forms that take
advantage of the possibilities of electronic media
both for
analysis and for presentation of the argument
, relying on
Geographic Information System (GIS) and on connect large
amounts
of information.view
which displays results extracted from that
database according to a query. The view component has gained huge importance so
that it usually influences the design of the data model.
In this section, we shall see how our process-oriented methodology can offer the opportunity to build a view component with the intention to visualize historical data.
We shall exemplify our methodology on case studies drawn from Ann Blair’s
We first chose
In chapter 2, the author sets the bases of a history of note-taking, as various kinds of writing in response to listening,
reading, or thinking…for the production of a composition or
report
erasable writing surfaces were commonly used for
temporary notes
physical
integration into a manuscript used for printing
as illustrated
through Pierre Bayle’s
taking notes,
discarding notesand
integrating noteswill be categorized as activities in the SIPOC representation whereas the so-called associated means will be represented by all kinds of
loose sheets stored in bundlesand notebooks as stipulated further. What could be the roles corresponding to those activities in the SIPOC flowchart? If we read Ann Blair more thoroughly, we learn that the notes taken by readers on erasable surfaces were often destroyed. Those
rolesin the
by epitome or abridgment,that is, by paraphrasing or summarizing the sources, or
by head or commonplaces,that is, by copying passages from the source into a notebook under a commonplace heading. These ways of working shed light on the
As the quantity of notes increased, it became of central importance to set up
tools for their management. One crucial tool was the assignment of headings to a
set of notes. As Blair mentions, the choice of such headings was rarely
discussed by scholars, although it was crucial to effective
note-taking
and required judgment, as part of a decision process
would be entered into the index under a keyword
under its principal thing
and to draw up an index on sheets corresponding to each letter of the alphabet
with references to the notebooks. That way, the index could receive new headings over time
. This indexation of a notebook is
therefore described as a living process which, together with the constitution of
headings, corresponds to an activity in a SIPOC that could be entitled
What was the main purpose of excerpting and accumulating notes in notebooks?
Blair tackles that question further in chapter 2. The historical materials she
presents in her development depict the notes taken by scholars as an aid to writing
, as worded by a subtitle of the
chapter. In our view, that
In chapter 4, more focused on the compiling activity, the author gives many hints
on the methods used by compilers and on the social and technical context they
were acting in. In particular, she discusses some social and legal elements,
which, to our view, should be included in the process description by a SIPOC.
For instance, the questions of obtaining privileges, which were only valid
within a certain jurisdiction
Regarding the compiling activity itself, it may be considered the result of
collaborations either diachronic, when compilers relied on earlier work
If we look more precisely at the technical activities as practiced by the actors,
we find some rather basic ones, such as selecting a passage from a source and assigning the
passage to a topical heading
the use of
slips and of cutting and pasting from printed books.
Thus, to describe an overall process, we can consider flow charts at several levels: at the top level, we represent global processes whereas more detailed processes may be specified at a second level by another flow chart.
This is described in figure 5 where we translated the note-taking activity according to the elements of Blair’s work previously reported.
As we can see in figure 5, each activity is still framed by a blue rectangle
which means that we consider possible another level of accurateness. Below, we
exemplify that third level flow chart with Sacchini’s
This foregoing review of Ann Blair’s chapters 2 and 4, translated into a process oriented representation, gave clarifications on the sequences of the activities (at different levels) that yielded a final intellectual product corresponding to the expectations of scholars in early modern Europe. But this generic representation based on SIPOC and flow charts only assigns a correct position to each component in the global process (roles, activities and associated means) provided that the component is effective in the process. If we focus now on a particular case with the intent of building its relevant SIPOC, we should be able to visualize its shift from the generic representation. In particular, we will inevitably have to contend with gaps in historical information, which amounts to the absence of certain components in the SIPOC flowchart. The choice of a particular case could be made in function of different criteria. For instance, we could follow the process of one scholar or a group of scholars who committed themselves to a particular production. That kind of process would spread over a certain geographic area, over a short period of time (that of the scholar’s career). In contrast, we could also take into account the achievement of a book over a long period of time, such as Theodor Zwinger’s
For the sake of the interruption in reading that resulted from stopping
to copy a passage into one’s notebook: it slowed down reading and aided
retention
in the order in which they were encountered
, and
second he copied the passages out in another notebook under commonplace headings
In this section, we showed that the process oriented view, which is the SIPOC diagram coupled with flow charts at different levels, proved to be a powerful candidate to visualize and represent data at the front-end of a digital system where historical work must be stored. Beyond its advantages in terms of visualization, it compels an accurate and rigorous analysis of activities, and in so doing, it enables setting up a powerful focus on specific cases. Therefore, it provides the scholars with a tool reference for comparing different cases and for raising subsequent questions. In addition, that tool does not concentrate only on individuals, but it defines, in a broad meaning, activities and roles related to each other, which gives an account of the interactions between the actors and their environment. Last, one more advantage of such a methodology lies in its ability to provide a global and overall representation of a past enterprise.
However, we must be careful when using this methodology. Indeed, the first problem we encounter when we try to apply the SIPOC and flow chart representation is related to the effort we must make to deeply understand all its subtleties. Indeed, this methodology was not first designed for historical works: it is then necessary to adjust it to the needs of historians (for instance, by redefining some of the vocabulary). Furthermore, it rapidly proves to be inoperative if the amount of scholarship evidence is not sufficient, as it is demanding many details to be enlightening.
This foregoing and detailed part has shown that the SIPOC-flowchart representation is well-adapted to a certain type of
scholarly work, namely any work related to the production of knowledge. But, we
can now legitimately examine whether that method of visualization could be
applied to other works. To answer this problem, we favored a pragmatic approach:
as our point is not to build a theoretical framework within which we would
demonstrate logically the validity and the relevancy of our method for a general
case, but rather to show that it can help visualize results of selected
scholarly works in a thorough way. To that end, we put our attention on a recent
publication,
textile copy and counterfeit in the Renaissance
As the reader can anticipate, this chapter of the book is rich and full of many details that can lead astray someone not familiar with that scholarly topic. After that first general presentation, the author expounds a detailed case study focused on linen production in Westphalia. As our purpose is not to summarize that fascinating study, we only would like to benchmark our method of representation to the previous general considerations.
In figure 7, we tried to render these considerations in a SIPOC-flowchart representation.
In this diagram, we chose to define six main activities controlled by four
milestones corresponding to the different inspections and certifications that we
briefly presented above. This flowchart rests on the principle that a process
which would successfully pass all these phases would produce certified and not
counterfeited fabric. All these activities are represented as general ones,
which means that they may be unfolded into subactivities to provide more
details. This could be achieved, for example, by extracting the information
contained in the footnotes of the chapter. Note that the bypass arrow between
the Dying quality control
and the Quality control
milestones
models a
For more than forty years, digital system architects or designers have adopted
common methods to build databases according to standard models. As previously
mentioned, an information system, that is to say the core of any Digital
Humanities project, usually rests on three main components (see figure 3). First, the data model, mainly a
database, contains the scholarship work organized according to a rationale.
Then, the view acts as an interactive component between the data model and the
end-user. It is a simple web page, a rich interface application (or RIA based on
JavaScript framework) or a traditional graphical user interface (java based for
example), depending on the level of sophistication. Last, the controller is the
component that connects to the database, retrieves the data relevant to the user
query and prepares them for a display on the view. The main advantage of such
architecture lies in its ability to distribute the overall development among
different specialized skills (database design, programming and view web-based
design). We have already presented the process map methodology as the view
component relevant to certain historical works. Let us now briefly describe what
could be another central brick in Digital Humanities: the organization of
historical evidence or the design of specific databases. In human sciences, the
databases must essentially follow the relational model based on the second
and/or third normal forms
Those technological “novelties” may not be regarded as simple tools by the
historians any longer. At least, they offer new visions on the historian works,
new ways of analyzing and new ways of researching flowcharts are good candidates for segmentation in atomic
entities as recommended by the relational second and third normal forms. Indeed,
we can anticipate, in the case of a database design, that activities would be
modeled by an entity which would cover several attributes such as its title, a
description and probably other useful fields. That entity would have a
relationship with others: a role
entity, an input
entity, an
output entity and a document entity. These relationships would be attributed
cardinalities whose values would depend on their nature. As it is obvious that a
role may be linked to several activities and, reciprocally, an activity may be
performed by several roles, the relationship activity-role will be of type
n-m
with a 0-N
cardinality on each side.input1
will be described
by first order notes, books, letters
. This is technically simpler
than considering three elements input1
(first order notes),
input2
(books), input3
(letters) that we should link to
the same activity through an n-m
relationship. It induces more
redundancy but offers more simplicity.
The diagram above must be made more complicated if we want to account for the
sequence of the activities, as it appears in the flowchart process. One way to model that sequence consists in
assigning a sequence
entity to the
subactivities, which amounts to create a
To practically exploit such a model it is necessary to convert it in a relational database (using the well-established rules), then to fill it with data. In order to represent a particular SIPOC, one could develop an engine which would retrieve the relevant data in function of a user request and which would generate an HTML page to display a flowchart. Indeed, the recent HTML5 standard includes a
One way to draw automatically a flowchart in
This data structure fits to the relational database model seen above and can thus be easily built in a client (browser). Then, in the client side, we only have to perform some appropriate
to create a SVG plane, and:
to create an activity block.
In figure 10, we reproduced the result obtained automatically by the code given in appendix. For the sake of clarity, we have represented only a few dimensions of the flowchart (in black arrows) relative to the web page edges.
In this article, we proposed a solution derived from a combination of several
best practices in order to account for scholarship evidence of professional
historical works. The best practices we are pointing at come either from the
industrial management process, from the software development and from the
critical distance that we must take to produce a relevant visualization of a
scholarly work. Concerning this latter point, we carried out tests of our
solution on two different topics that were objects of investigations by
historians, namely the production of knowledge and the question of
counterfeiting in the Renaissance. We saw that the SIPOC-flowchart representation gives an overall view of a case study that
resulted from a scholarly work. It gives an insight to the shape and structure
that constrain the practices of a community. But we also argued that building a
SIPOC-flowchart helped us raise problems and
questions that we did not anticipate before, in the sense that our
representation suggests (through the blanks that it left) solving and answering
particular problems.
Last, we think that it is very likely that any other work involving process or,
more generally speaking, solidarity between different actors in a specific
social context would be easily translated into the Digital Humanities field by
our solution. Indeed and in a practical way, once we have properly designed a
database to represent data coming from historical works, we open the door to the
web ecosystem and to all the features it offers. In particular, we can imagine,
following Moretti’s call for a worldwide collaboration