Friday, January 25, 2008

The Attali Report in Toto

If you care to read the Attali Report, it can be found here.

An article on calls attention to the report's emphasis on promoting information technology and singles out articles 53, 54, and 58 for special attention. Article 58 advocates promoting open source software as a competitor to proprietary software and specifically calls for greater use of open source software in the public sector. "A goal of 20 percent of newly developed or installed open-source applications for the benefit of the public sector could be set for the year 2012."

The Attali Report has been characterized as "liberal," but this can hardly be called a liberal measure. It sets a fixed quota for the use (or is it the new development--the wording is hardly a model of clarity) of open-source software, which would oblige public sector organizations to procure from a specific source without regard to the competitive quality. To be sure, the subsidization of open source (envisioned in the same article in the form of a tax subsidy) could be interpreted as a blow against the "evil empire" of Microsoft, which in French eyes combines the sins of monopoly and American nationality. Nevertheless, I'm not sure that it's a wise idea to enforce by fiat the choice of a software regime, and I say this en toute connaissance de cause: I am sitting between a Windows machine and a Linux machine, and there is a Mac in the other room. These peacefully coexist in my household, but I have a Ph.D. from MIT (honest). Others should proceed with caution and certainly avoid the mistake of deciding that within 4 years, "twenty percent of the software" in this or that office will be open source. What does that formulation even mean? If I have an (open source) Apache Web server installed on one machine and Microsoft Word on 4 others, do I meet the 20 percent requirement? Is it OK if I execute a gazillion instructions a year under (open source) Linux and 4 gazillion on Mac OS X?

This ambiguity is typical of the Attali report (or at any rate, as much of it as I have read). Its recommendations suffer from being neither general nor particular. They create a false impression of specificity by recourse to arbitrary numerical quotas and technical jargon (viz., art. 53: "establish a European mechanism of digital identification allowing a mutual recognition of means of authentication by requiring root certificates issued by European certification authorities for the entire suite of communication software [messaging, browser, etc.] sold in Europe"). This is not the level at which such a commission should be operating. The pseudo-specificity is just eyewash. It may enable Jacques Attali to pose as an expert on everything, but in reality he is just an expert poseur.


Unknown said...

"Expert poseur" That's really excellent and so relevant.

Anonymous said...

Excellent point on the metrics, which seems from here like it might part of the Sarkozian mania for numerical targets generally. (Which can be dangerous even if the numbers are well defined: if you reward computer programmer productivity as measured in lines produced per day, you're rewarding code bloat and penalizing elegant design. And that's before some ministerial Kerviel starts gaming the numbers).

I'm not sure it's fair, though, to call the entire open source community a single, "specific source". If you want to use open source products to, say, develop a web site, you might use Mysql or Postgres (or others) as the database; Rails, Django or Struts (or others) for the application logic tier, Apache or Lighttpd (or others) for front-end web service; and run the whole thing on Linux, OpenBSD, Solaris (now fully open source)... or others. And that's before even getting into the multiple options you have for commercial support with a lot of these packages.

Agreed, though, that the numeric target isn't much use, and that more specific, concrete proposals would be useful. (It would be nice to see something, for instance, on limiting the activities of patent trolls, which wasn't obvious to me on at least a quick skim of the relevant sections of the report.)

Anonymous said...

Art - You're right about Attali - and he's been an expert poseur for a very long time. I met him originally in 1970 when Annie Kriegel brought him to a seminar at Nanterre to discuss his first opus, a "Que sais-je", if I remember correctly, called "Les modeles politiques," a kind of rat choice avant la lettre(I was a poor graduate student adjunct hire of Annie's for one class on methodology, about which I knew more or less nothing, not even a poseur. It didn't matter because a la rentree Nanterre was shut down by student/faculty protests. I thus began my career on strike.)
Attali is larded with talents -but he's spread himself too thin and followed his ambition to try his hand at almost everything: including novels (unreadable), conducting a symphony orchestra (n'en parlons pas)and creating the BERD (not such a bad idea).
He milked his Elysee office -- the necessary transit to Mitterrand's -- for all it was worth.
I never knew him well but my impression of him -- always insufferable but in some ways genuinely admirable -- is probably the consensus view. He's had a good run.

Nora von Ingersleben, ACT said...

While I think that the Attali Report contains quite a few sensible and useful recommendations (such as concentrating tax credits for R&D expenses on key high tech sectors, including the digital technology industry, and reducing the payment time for sums due from the state and major enterprises to SMEs), I agree that the report's recommendations on open source are completely misguided.

Telling the government that it has to use (let's assume Attali means "use") 20 % open source software by 2012 amounts to giving an advantage to open source through public procurement and, as you rightly point out in your post, undermines state agencies’ flexibility to choose the technology that best suits the needs of the task at hand.

By the way, AFDEL, the French Association of Software Developers, shares our concerns about the “20% open source objective.” In a Reuters article and in an article in French business daily Les Echos, AFDEL stated that the recommendations on open source “undermine investment in innovation.”

Given that French politicians have largely ignored the many reports produced over the last few years that prescribe a cure for the ills of France's economy (there was a good article on this in the January 26-February 1 Economist), I am not sure how many of the recommendations contained in the Attali Report will actually be implemented.

But should President Sarkozy decide to adopt some of Attali's suggestions (and he has already said that he won't adopt all of them), let's hope he will steer cleer of implenting the report's recommendations on open source.

(If you are interested in reading my blog post on the Attali Report, you can find it at