Introduction

In this text, which follows on from the first part on Microsoft, we show that the extreme difficulty of challenging Microsoft’s monopoly through free competition has prompted a reaction from society under the banner of the middle class. This is the phenomenon of libre software. Its relative success in the context of the Linux operating system, was then relayed by the bourgeoisie and its states to give it a coherent economic foundation. The text shows that this response is in no way socialist.

Impossible market response

We have shown that, from the point of view of capitalist competition, we can expect nothing from an equivalent mobilization of social work on the part of other software manufacturers to counterbalance Microsoft’s influence.

So, on the one hand, we have the liberal bourgeois response based on free competition and anti-trust. This response remains purely ideological, because, as we have shown, it would not have the desired effect on the general level of software prices. On the contrary, it would require a much greater expenditure on social work to achieve similar results. This response is therefore reactionary.

More interesting than the bourgeoisie’s response to Microsoft is the reaction to libre or open software – both are not the same thing - whose emblematic figure is the Linux operating system.

A roundabout response initiated by the middle class

This response originated not in the bourgeoisie, whose responses are inadequate by the very nature of the software product, but in the middle class.

If the Western world telephones or turns on the television without too many complications (it’s already more difficult when it comes to programming the VCR), and without any particular knowledge of the technologies specific to telephony or radio broadcasting, more sustained use of a computer - much to the chagrin of many users who don’t want to be burdened with this knowledge - regularly leads to the need to take an interest in questions relating to file formats, printer settings and certain operating system elements, the subtleties of application software, and so on. As a complex object, the computer requires the disproportionate spread of a computer culture throughout society if it is to be used effectively. The need to produce software, as well as to parameterize certain software packages to adapt them to specific needs, has also led to the development of a programming culture. This culture is all the more readily accepted as it enables its owner to carry out both intellectual and material work, to produce a direct effect via the machine he or she uses. Rapidly controllable, modifiable at will, without any significant physical effort, without getting your hands dirty or bothered by dust, with a right to make mistakes that is all the greater since you can erase everything and start again without expending many resources other than your working time, programming has won over a large number of enthusiasts.

A significant potential for software development capacity, which results in particular from the complexity of computers but also from the need to adapt software needs as closely as possible, has therefore been organized in society. This is particularly the case in research activities.

This is the breeding ground for a specific reaction to Microsoft’s monopoly - a reaction marked by the stamp of the middle class. Researchers, academics, executives of large public companies and administrations, independent developers, students, employees of software companies who redo in the evening what they do during the day, and young enthusiasts are the main vectors of libre software.

They have drawn on their opposition to Microsoft, their technical perfectionism, their desire to promote and train, and their desire to regain control over their tools (Linux is not always easy to install and configure for the uninitiated), to launch an assault on the monopoly. They have also found a type of product: software, and an infrastructure: the network and the Internet, which give them the right foundations to recreate a particular form of associated work.

A middle-class anti-proletarian, anti-communist ideology

In his reply to Raoul, JC highlighted the anti-communist ideology of libre or open software, so we won’t dwell on this point.

[[The correspondence referred to in this section is lost]]

The fuss of the bourgeoisie

It’s interesting to note that we’ve moved on from libre software to open software, i.e. software whose source code is available and can be modified by its user [1]. Open software is not necessarily libre, and this semantic shift is a manifestation of the fuss thrown up by the Linux initiative among the bourgeoisie, who saw it as an opportunity to beat Microsoft to the punch, once the product had proved its worth.

This attention from the bourgeoisie took various forms. It ranged from companies who saw the opportunity for new types of market based on open software, to major manufacturers like IBM (whose sales are greater than Microsoft’s) seeking revenge.

On the other hand, in the open software economy, while the license may be free (this is not systematic for open software), the same does not necessarily apply to implementation, parameterization, adaptation to the user’s specific needs, training, maintenance and so on. The “experts”, whose credentials are not necessarily high, estimate the gains on the entire software chain at 40% of the total cost. Of course, this is not negligible, and can be a godsend for many companies and public authorities. For supplier companies, this chain of activity is also likely to generate profits.

Another aspect of this fuss is worth highlighting here. Companies involved in open software tend to dedicate employees or, more modestly, provide the results of developments to the open software community. This is by no means a philanthropic action, but here the social work - unlike private volunteer work - is mutualized among the various users. The company does not charge its customers directly for this development work, but indirectly (in the same way, it indirectly charges its customers for commercial work - not only the commercial work required for the market it has just won, but also for all those it has lost).

Technical progress = progress in the exploitation of the proletariat

Insofar as it is a reality, the fall in the price of software applications will translate into a fall in the price of the elements of constant capital, and therefore a fall in the price of commodities. As a result, the value of labor power will fall. The savings generated by the fact that part of private labor, destined to become free social labor, is made available to society, contribute to the increase in surplus value. One part of society gives its labor free of charge to another part, in this case capital, which gains in value from the increased exploitation of labor power.

Free labor and socialism

Raoul tells us that free software developers are all the more gratified because, with less effort (the software is only developed once), they multiply their possibility for recognition (the software can be duplicated ad infinitum at a relatively low cost) [2]. So be it! [3] but the logic of maximum gratification for minimum time is in no way exemplary, and is merely the continuation of the capitalist logic of maximum result for minimum effort.

Raoul is fascinated by the phenomenon of gratuity. But this is in no way specific to the world of libre software. Let’s disregard radio and, to a certain extent, television programs, certain newspapers and some corporate activities - such as drawing up an estimate - and look at other forms.

From the person who bakes a cake for his neighbor to the thousands of people who get involved in various associations, they all have a similar approach.

Every day, thousands of volunteers give their time to a wide range of associations. They give their time with, no doubt, far less ulterior motive than developers, and without seeking as much recognition. What’s more, whereas development is a one-off event, with the possibility of general dissemination earning them the gratitude of many as well as the recognition of their peers, associative action is renewed on a regular basis.

There’s at least one other area where free labor plays a considerable role, and that’s in the family. The lion’s share of this work is done by women. It’s possible that Raoul sees in these “seeds of non-market relations” the pre-homo socialisticus, but revolutionary socialism has so far seen in them only the mark of the exploitation of women.

Raoul imagines that software has a specificity, however: because the fact of duplicating it for practically nothing does not lead you to divest yourself of it, and that it is a means of production and also of consumption.

The results of science are in the same situation. It may have taken a scientist a lifetime to find this or that theorem, but it only takes a teacher a handful of hours to communicate it to thirty pupils. Afterwards, as before, it remains in the teacher’s head [4]. As science is incorporated into production, it obviously becomes a means of production. Every engineer has to make calculations, work out rules of three or use the Pythagorean theorem, to take trivial examples. The computer is a machine capable of automatically processing algorithms. But algorithms exist independently and predate the computer.

When Raoul enthuses about the potential effects of free software, apart from totally overestimating the importance of software in the cost chain and mistaking petty-bourgeois bladders for proletarian lanterns, he is merely describing - and this is the only grain of truth in his discourse - the dissolving effects of the progress of the productive force of labor on capitalist relations of production.

Communism or petty-bourgeois socialism

Insofar as these activities are part of a process of challenging bourgeois society and the way it functions, the Communist Party has long qualified them as varieties of petty-bourgeois socialism or bourgeois socialism. Consult the Manifesto of the Communist Party.

Socialism does not consist in everyone baking a cake for their neighbor, free of charge, in their own kitchen. It is not a generalization of the gift economy or charity.

Nor does it mean that, within the framework of corporate socialism, workers’ councils share the fruits of labor and that each worker goes home with cakes in his bag.

It consists, on the contrary, in placing the most developed productive forces under the control of associated producers, breaking down the limits of the enterprise, reorienting and reorganizing the productive forces in such a way as to reduce the necessary work to its simplest expression towards the satisfaction of human needs, while improving the sanitary, alimentary and gustatory qualities of the cake; part of the vestiges of the enterprise being transformed into simple production sites.

Is the petit-bourgeois solution the right one?

Although imbued with a petty-bourgeois spirit, the fact remains that the development method worked, even if this success was aided by the bourgeoisie. This form of collaborative work is not social work in the strict sense of the term, but becomes so in a particular form, in the sense that it becomes free, via the mediations we described above. Compared with the organization proposed by socialism, based on associated work, is this form of organization more efficient, less time-consuming?

To answer these questions, we need to turn to the actors and witnesses of this type of organization.

A key witness

Eric Raymond’s article “The Cathedral and the Bazaar” is an emblematic text from the Linux world. Eric Raymond is an experienced programmer and the figurehead of open-source software, where Richard Stallman, President of the Free Software Foundation, is the figurehead of libre software.

Eric Raymond was already heavily involved in the world of libre software back in the 80s. He believed, however, that beyond a certain level of complexity, it was not possible to develop software in a decentralized way, with thousands of people working together in their spare time, which is what made Linux such a success. He also believed that it was necessary to reach a certain degree of software advancement before releasing an experimental version to advanced users.

The cathedral and the bazaar

Eric Raymond thus contrasts the development style that Linux would have inaugurated via Linus Torvalds - rapid and frequent distribution, massive delegation, total openness - the bazaar style [5] as opposed to the cathedral style described above [6].

What’s more, far from collapsing in confusion, the bazaar style was progressing by leaps and bounds.

Believing he had understood the methodological foundations of the bazaar style, in 1996 the author tried to put it into practice on one of his projects. The text “The Cathedral and the Bazaar” recounts the adventures of this project.

Social need

Faced with an e-mail problem (a personal one), the author drew a first lesson from the experience: good software corresponds to a developer’s need (all good software starts by scratching where something doesn’t work).

According to the author, this is the case with software from the Linux community. Generally speaking, we’re going to see the emergence of the point of view of a technical aristocracy, a technician caste, which is at the root of the Linux ideology, and which more broadly corresponds to the ideology of the middle class, the petty bourgeoisie.

This is one of the difficulties and limitations for companies. If Linux-type software development is dependent on the fact that it must respond not to a social need, but to the particular need of the developer, then social development is dependent on the particular need, and only expresses itself insofar as it meets the particular need. Only if the particular need meets the social need can it be expressed. Obviously, a completely different logic applies to the production of software for the market. In this case, for the private work included in the software to become social work, it must be recognized on the market and therefore meet the social need. It’s then up to the company, via multiple channels: sales feedback, market and consumer research, customer requests, complaints and after-sales management, competitive and technological intelligence, to amend the product so that it is always better adapted to this need.

As a result, Linux-type development will meet the social need all the better if it is confined to developers’ issues (operating system, development utilities). Consequently, the likelihood of such software being located upstream in the software chain will be all the greater.

The article closes in 1998, with the prospect of developing a Netscape browser using the same methods assumed by the Linux community. A few years later, it has to be said that this software has not only failed to establish itself, but has regressed in its social use, to the benefit of Microsoft’s browser. [[1]]

Intergenerational collaboration

Then there’s another trait of this technician aristocracy: the idea that it’s laziness that makes software development progress. “An important characteristic of great programmers is constructive laziness. They know that you don’t get 20/20 for effort, but for results, and that it’s almost always easier to start with a good partial solution than with nothing at all” (Eric Raymond, p.3).

As we have already mentioned at length, this development has particular economic characteristics, since the time required to reproduce it is divorced in proportion to the time required to produce it. Added to this is the fact that it is a task that implies a strong incorporation of science into the machine. It is both intellectual work and physical work that enables the machine to function, to provide the service for which it is intended.

On the other hand, as an intellectual product, it is particularly flexible, since lines of code can be repeated. In fact, this means that we are building on the work of previous generations, that in this type of work there is inter-generation and intra-generation collaboration, that the software is not the work of an individual, even if the illusion remains. We are not dealing here with the work of a writer who has to write from the first to the last line, but who benefits from the state of knowledge on all levels of his era, nor with that of a scientist who inherits the theories, laws and formalizations of others in order to apply them or develop others, but makes use of a form of collective writing that enables him to delude himself about his individual prowess while still being able to achieve tangible results.

What we’re saying here needs to be qualified by the fact that, in these communities, there’s a temptation to consider that what isn’t done by the developer himself is no good, and it is common to have the greatest difficulty in maintaining another program, which is generally poorly documented due to the lack of time and the developers’ lack of interest in a task considered to be boring.

That said, Eric Raymond points out that in the world of libre software, there is a huge availability of open software, which makes it relatively easy to find software that already has many of the features you’re looking for. In this sense, he shows the limits reached by the capitalist mode of production and the division of productive activity into the autonomous cells that make up companies. The sharing of universal labor (the intellectual labor of past generations) is all the more hampered the greater the competitive universe. The necessary overcoming of the competitive and mercantile economy (symbolized, in a limited way and with its known shortcomings, by the monopoly) is therefore inscribed in the progress of the productive force of labor.

Associated labor

Many of Eric Raymond’s other considerations are a magnificent blend of common sense and naïveté, flavored with Anglo-Saxon morality. We’ll just mention the idea that associated work is superior to individual work, which is a truism for socialists, but not for the individualist community of programmers.

We’ve already highlighted the limits that competition and capitalist production place on the generalization of this possibility, and how it finds its way in through the concentration and centralization of capital, monopoly, nationalization, multinational or transnational firms, Economic Interest Groupings and other inter-company projects that lead to the overcoming of these limits without ever completely succeeding. It remains to be seen how this cooperation can be achieved within a framework that is neither the market nor the associated labor characteristic of a communist society. This brings us back to our subsidiary question. Is this cooperation superior to that advocated by socialism on the basis of associated labor?

For Eric Raymond, this is the secret of Linux: the ability to turn user-programmers into co-developers.

In the world of open software, this possibility has been put to best use in the production of Linux. One of the new paradigms, according to the author, is that of not hesitating to release new software updates quickly, without worrying about possible bugs, while at the same time listening to users (at certain times, the Linux kernel was updated several times a day), all served by an ad hoc infrastructure, namely the Internet.

So what Linus Torvalds is said to have done is not the work of a genius designer, but “Linus constantly stimulated and rewarded his users/tinkers. He stimulated them with the self-gratifying prospect of taking part in the action, and he rewarded them with the constant (even daily) sight of improvements to their work. Linus sought directly to maximize the number of person-hours thrown into the battle of debugging and development, at the possible cost of some instability in the code and the extinction of his user base if some serious bug proved insurmountable. (…). Given enough observers, all bugs are obvious; that’s what I call Linus’s law” (Eric Raymond p.6) (7) The above remarks can be rephrased as follows: “debugging can be parallelized”. (Eric Raymond. p.7)

But Raymond has not shown that this does not result in a significant waste of work time. Nor does he show that the social cost generated by “a thousand greedy co-developers” who “rush to any new update” is any higher. Admittedly, this makes debugging progress very quickly, since the more users a piece of software has, the more efficient debugging is, but it doesn’t show that the marginal cost of debugging is lower, and it probably isn’t. It isn’t because the more users it has, the more efficient it is. It isn’t because the greater the number of users, the greater the number of bugs discovered, but that’s not the point. With a comparable scope, the time spent will be greater than with a centralized approach. A swarm outside the hive doesn’t make more honey in the same amount of time.

It could be argued that software quality improves more quickly, but in a way, it is neither possible nor desirable to follow this debugging method. In this case, when a user discovers a bug and reports it, he loses time, but it’s marginal time. If it’s counted in the social cost, the time spent will be much greater. But as it’s a task that doesn’t form part of his traditional activity, this time is rarely taken into account in evaluations. It’s free time given by the firm, but the firm would have to spend it anyway if it wanted to eliminate a bug, by artificially reproducing the software’s conditions of use. As a result, all software is more or less buggy when it arrives on the market. In the case of Linux, as users are also developers, they are also in a position to propose debugging solutions (and generally not for the bug discovered). What the story doesn’t say is under what conditions do they discover bugs? Are they acting as users in the full sense of the term, or in part as experimenters who, because of their interest in the product, for ideological reasons - notably hostility to Microsoft linked to the libre software milieu - or for professional or educational reasons, would make their work available to development free of charge? How much of this work can be likened to dedicated work, and how much is marginal? The greater the amount of dedicated work, the more we have to conclude that, from a social point of view, the way in which a product like Linux is produced implies a major waste of private work - a waste that the community formed by a technician aristocracy is prepared to make, for various reasons already mentioned.

Conclusion

The bourgeois response to Microsoft’s monopoly is ineffective and reactionary. To be effective, it had to rally to the middle-class response and give it a credible foundation.

The middle-class response, namely libre software, is an effective response. It consists in opposing social work with a multitude of private-time jobs, and making this private work freely available to society, i.e. to capital, thereby socializing it.

Insofar as it is genuinely effective, which is partly the case, this response enables capital to increase its exploitation of the proletariat by lowering the value of goods and services that use this free private labor in their production. On the other hand, this response is both utopian and reactionary. It presupposes an immense expenditure of private and generally isolated labor to counterbalance the efficiency of associated labor as implemented in companies.

So, although it is, like all middle-class ideologies, both utopian and reactionary, the prospect of libre software and, more precisely, of open source software, has received the support of the bourgeoisie for want of any other response. Governments and major manufacturers such as IBM have sang the praises of the middle-class initiative to give it greater overall coherence, so that Linux now appears to be Microsoft’s most serious competitor. At the same time, the ideal of the technically-minded petty bourgeoisie withered in the face of the mercantile temptations that accompanied this relay and institutionalization of open software.

The proletariat’s response, totally absent from the scene, and based on the socialization of production, the collaboration of associated labor, and the break with the corporate economy, is the only consistent theoretical and practical response, the only one that liberates the productive forces of labor in the service of humanity. It is the only revolutionary response.

Robin Goodfellow - May 2003

NOTES:

1. This argument is particularly debatable in the case of the operating system. If anyone can modify it, the Tower of Babel is guaranteed. In fact, it’s well standardized and commercialized versions of Linux that are installed (e.g. Red Hat Linux, named after the company that maintains the versions).

2. However, let’s not forget the costs involved in installation, training and maintenance, which are nowhere near the cost of reproduction.

3. The psychology of developers is slightly different. It’s not so much a question of producer-user recognition, even if this dimension does exist, but of a more complex mix of recognition by peers within a technician caste both and primarily from the angle of technical expertise and the aura it induces in the technician community, and from the angle of the user - the users who come forward to the producer often being members of the caste themselves -.

4. This kind of consideration is taken up by any charlatan dealing with the information economy. In an exchange of ideas, each party takes away the other’s ideas without having abandoned its own.

5. It’s been tempting to call it Agora-style.

6. It’s a misnomer for JC to say that the cathedral style is aimed at Microsoft. Eric Raymond’s text refers to an organizational guru, P. Brooks, author of the myth of the man-month, for whom bazaar-like practices can be attributed to Microsoft. (p. 18)

7. Raymond also makes the interesting observation that the person who identifies the problem and the one who finds the solution are not usually the same.

TRANSLATOR NOTE: [[1]]. This is quite ironic, given the scenario has almost changed, and repeated itself once more. Libre Netscape turned into Firefox, which, for a brief period, looked like it had the chance to dominate the browser market. However, Google Chrome has come around in recent years, and has once again claimed first place. Even other browsers use Chromium as their backbone.