Chapter 15
The technological revolution of the digital and robotics

In chapter 6, we saw that the two most specific elements of our era are, on one side, Earth suddenly becoming a limiting factor in our development, giving rise to the ecological issue, and on the other side, the second industrial revolution, that of the digital and robotics. The challenge for our generation is therefore naturally to adapt our social organization to this new environment.
We just saw in the previous chapter that the ecological transition implies moving away from the logic of growth. We are now going to show that the digital revolution implies reconsidering digital tools and training, in order to genuinely make them available for the operational organization of work.

Emergence of a new tool of oppression

Marx noticed that when machines arrived in workshops, rather than being liberating for workers as one might have expected, it resulted in greater stress at work. Exactly the same thing happened with computing. Why? Because the massive investments, and hence the focus of all capitalist attention, went to the machine aiming to replace muscle in Marx's time, to the information system aiming to replace the brain in ours. People are viewed in both eras as simple adjustment variables waiting for future automation. History repeats itself at this level.
However, if capitalists have approached the technological revolution of the time in the same way in both eras, the result was not the same. In 1987, Robert Solow was surprised that the massive arrival of computers in companies at the beginning of the 1980s was not reflected in productivity gains statistics, which will later be called the Solow paradox. The explanation we find most credible is that if finding a good organization around mechanical machines turned out to be reasonably simple, it is not the case around computer systems. Indeed, with mechanics or paper bureaucracy, the simple visual observation of the flow of the goods or forms allows to get a good idea of the current level of organization, and to imagine improvements to fluidify production. With computing, it is much more abstract because you don't see directly the progression of the files. You need to build a mental image from simple indicators. Moreover, when elaborating an improvement to fluidify or simplify production, it is extremely difficult to assess the level of difficulty related to the implementation of each option.

In fact, for now, computing has been an accelerator of the problems described in chapters 2 and 3, namely the implementation of a formal organization disconnected from reality, as described in the article of Meyer and Rowan, and the inflation of non-productive prestigious staff, as described in Parkinson's law.
Indeed, in practice, computing has introduced a third organization system, in the form of generic activity modeled in computer software, which adds to the formal system described by Meyer and Rowan, next to the informal organization which ensures the actual production. However, the emergence of this third system causes operational people to lose their autonomy and become unable to coordinate naturally and smoothly as they did before, simply because the necessary computer changes do not follow or are simply not feasible. This leads to a proliferation of advisors of all kinds, which is ultimately the acceleration of Parkinson's law.
This phenomenon is further amplified by recent technological evolution, namely the development of instant communication means such as email then mobile phone with its short messages and other collaborative tools, which make people reachable anywhere and anytime. Indeed, the fact that information circulates faster has favored the implementation of increasingly long chains, thus constituting another powerful factor favoring the aggravation of what Parkinson describes, namely the inflation of non-really productive positions, where one simply dispatches the real work.

On the organizational level, computing has been an accelerator of the slicing of the company into silos, since each new software creates a new silo, which contains only part of the information necessary for the functioning of the organization, and guards it more or less jealously, as any problem journal quickly reveals.
Once again, web technology, with the emergence of multiple online services, has come to worsen the situation. Indeed, companies, facing the drift of their organization due to inadequate computing, conclude en masse, following analyses that do not take into account the sociological dimension, that it is their internal computing that they do not manage well, or which is too costly. They then try to outsource it to the cloud. The effect is a greater fragmentation of data, hence an aggravation of inconsistency problems as the number of services used increases, with the final consequence of a new acceleration of Parkinson's law, that is, the proliferation of tasks without direct link to the organization's purpose. Computer software, cloud or not, is a bit like medicines: too many medicines end up poisoning the patient because of poorly mastered drug interactions.

Finally, at the inter-company level, each company or administration seeks to set up a website on which they encourage or impose on their customers, suppliers, and citizens to enter or retrieve information (1). The local effect, for the company or administration, is to transfer the administrative cost to the outside, without transferring the corresponding remuneration, so it is extremely motivating. However, at the community level, the net effect is a considerable loss of productivity. Indeed, on one hand, each company, instead of having to manage a uniform administrative flow corresponding to its activity, ends up having to manage multiple micro-administrative flows corresponding to each of its customers and suppliers, with the impossibility of optimizing anything due to heterogeneity. On the other hand, as soon as the company or administration has outsourced - distributed on its customers and suppliers - the administrative cost, which has become painless for it, nothing limits its propensity to ask for more information, to feed its own multiple non-production levels, which can proliferate all the more easily.

Let us recall at this stage that an uncontrolled Parkinson's law translates into practice by an overabundance of management who intensely practice generalized nepotism, and pressured operational staff, therefore a high level of stress everywhere, that is to say, progress that eventually leads to oppression.

The choice of the computer system

We saw in chapter 4 that strategic choices concerning the computer system are among the most glaring examples of weak or inept reasoning. We mentioned the massive recourse to social support, to which is added the cognitive bias of overconfidence seen in chapter 3, as well as the asymmetry of risk-taking before and after the initial choice.
However, the consequences of these choices are both very important, not only because of the initial investment related to the development or acquisition and implementation of an integrated management software, but also and especially because of the effect on the general organization, and therefore ultimately on the social climate. Very concretely, if too many problems lead to the need to adapt the computer system, then the adaptations are not made, and constructive exchanges stop in favor of a simple power relationship, that is to say, generalized nepotism.
Paradoxically, unionism has not grasped this issue. There is no quality requirement towards management at the decision-making process level, and simply confrontation once the problems are there... and it is too late.

This means that computer strategic choices must be more than any others submitted to a rigorous methodological evaluation, as proposed in the second part of this book, simply because their consequences are heavy and not seriously anticipated at the moment.

A computer system under control

Let us now move to the solution. The previous chapters have shown how to reunify the formal organization with the practical organization by adopting a structure that ensures that the actual organization remains connected to the purpose, and is optimized according to the practical problems encountered. However, the search for the causes of the acceleration of Parkinson's law under the effect of the arrival of computing has just shown us that the source was always the loss by the operational staff of their ability to modify smoothly and continuously their organization. Let us see how this very concretely affects the functioning of the new organizations we presented in the second part of this book.
In the problem journal (chapter 9), at the 'Solution' level, what is important is that the weight of the modifications to be made in the computer system, and the difficulty to access the complementary information needed to prevent the problem from recurring, do not make the obvious solution of adapting the computer system almost never applicable in practice.
Similarly, in the strategic reflections journal (chapter 10), a major issue is that the constraint of computer systems that cannot be practically adapted to the need does not become what leads strategic reflections to become mere searches for tricks without long-term sustainability.
In other words, the danger is that computing produces the same effects as a fussy administration. Hence the importance of the 'Restoration of margins of maneuver' paragraph in the same chapter 10.

Let us therefore now study what is currently opposing operational staff to retain control of their computer tool.
First, on the technical level, computer tools have diverged, either downwards or upwards, making them in all cases unsuitable for direct and effective use by the operational staff. Downwards first, we have seen a sort of gold rush towards mass-market tools for consumers, whether web services, messaging and other social tools, which have become intuitive and largely mastered, but are not production tools. Upwards, the development tools have become significantly more complex, making them unusable by non-experts. Finally, in the middle, the business software present the same adaptability limits as the mass-market software, without keeping the intuitiveness. The cause is that decision-making buyers still believe that the quality of software is linked to the number of features, instead of understanding that the real quality lies in the possibility of adapting it simply when the problem journal brings up inadequacies with the actual activity.
On the training level, we have simply abandoned the goal of training a digital upright citizen in favor of an employability goal. As a result, training has followed the divergence of tools towards the bottom and the top. Downwards, we train in large numbers to tools such as word processing and spreadsheets, which do not allow setting up an efficient organization. Upwards, we continue to initiate programming in classical languages, which will not provide effective autonomy to non-experts.

The solution is therefore the adoption of a medium technology, which allows automation while remaining manageable by properly trained operational staff. This implies moving towards a homogeneous computer system rather than trying to interconnect multiple specialized software, so that the automatic grouping of all necessary information for each activity remains easy. In other words, a good computing system is nothing less than a computing system that does not oppose the good functioning of organizations, and in particular of the problem journal and strategic reflections, and this simply implies selecting technologies adapted for this purpose, instead of selecting technologies adapted for isolated computer development from production.

At the inter-company level now, it is necessary to define a single exchange document format (2), understandable by the digital upright citizen, such that a digital flow is rigorously equivalent to a paper flow.
Furthermore, it should be imposed that any operation possible via access to the company, administration, or organization website, is also possible automatically via the sending or downloading of some clearly accessible digital forms.
The goal here is that the boundaries between organizations do not become the areas where the work foreign to the organization's purpose proliferates, because the complexity and unnecessary heterogeneity prevent operational staff from effectively running the continuous improvement principle we presented in the second part of this book.

In the end, what the Solow paradox teaches us is that in the case of the digital, what is humanly desirable in terms of organization, because it favors the autonomy and responsibility of the employees, is also desirable from the point of view of pure operational performance, because it is one of the conditions for an effective problem journal.

Digital Policy

America, being the leader of mass-market computing services, logically experienced this gold rush by looking more at commercial profits than at the limits in terms of productivity gains missed in other sectors. However, Europe remained insignificant on the digital level simply because it did not know how to promote products adapted to operational upright citizens. It satisfied itself, on one hand, by trying to copy the Silicon Valley startup model, forgetting that copying with fewer means is rarely a winning strategy, and on the other hand, it satisfied itself by wanting to regulate via law, that is to say, go back to the method suggested by Marx in Capital.
Cognitive dissonance did the rest by spreading the idea that communicating more and faster can dispense with thinking about the organization, with the final result of a policy insignificant from the point of view of regulating the pressure on the operational staff, and therefore a return to social confrontations.

The key to the digital challenge is therefore a resolute policy of training to get out of digital illiteracy, and this implies selecting technologies and protocols adapted to the digital upright citizen, instead of continuing to adopt the technologies already selected by the market, that is to say, targeting either the bottom the digital consumer or the top the digital specialist, with the final effect in all cases the aggravation of Parkinson's law and the stress level that follows.

(1)
Many companies stop using paper to send documents, for example invoices. Instead, they send an email such as 'We inform you that we have made available such and such a document on your personal space.' Translated into plain language, this means 'Come and get it.'.
The fact that the legislator does not call to order the organizations that, in the course of going digital, cease to fulfill their obligation to send an invoice, is representative of the effects of digital illiteracy: since the lawyers do not master the digital, they under-invest in it, and it tends to become a field where might makes right.

(2)
Every form must be available in readable format, for example a PDF file. It is exclusively made up of elementary fields and tables. Each elementary field has a name. Each table has a name. Each column of each table has a name. All these names are unique and must appear in small text just next to the relevant fields.
The exact JSON encoding of the document is established canonically from the names appearing on the readable version. The details of the specification of this canonical JSON format go beyond the scope of this book. Here, only the principle of such a format interests us. Moreover, for greater clarity, an example file should be made available in addition to the readable version.
Regarding transmission, in an ideal world, the readable form would indicate the address of the mailbox or deposit it in digital form. However, due to technical limitations, we prefer to indicate a URL (a web address), and the document will be deposited by means of an HTTPs POST request, with a basic authentication of user and password, and return an acknowledgment document. In case of rejection, the HTTPs error message should indicate the name of one of the problematic fields, or, if it is a field in a table, the triplet table name, line code, column name.
For access to information, it is almost the same: we send a form specifying the requested information, and we receive back not an acknowledgment but a JSON response form.
Finally, it is up to the community to standardize a certain number of these forms, that is to say, precisely define the fields and their names, to further facilitate exchanges.