Some thoughts on the Conseil National du Numérique interoperability study

Some gems from the French national digital council study on interoperability:

The French Treasury is also in favour of proactive, proportionate and targeted regulation on structuring platforms, such as the obligation to develop technical standards facilitating #interoperability of services and possibilities for user migration.”

Structuring is “based on quant. criteria such as no. of users, % of market, or $$ of the platform, but also more qual. such as access or possession of quality data, or the control exercised by the platform with regard to its users (captivity) or 3rd-party companies (addiction).”

German draft law: “‘of major importance for comp. b/w markets”, … its dominant position…, €€, its vertical integration or diversification into ancillary markets, its access to data critical to competition, or the decisive nature of its services for … other market players.”

For some interviewees, “this power of constraint would be precisely the fact that we cannot leave a social network without losing the links we created there. It is therefore the actors who have such power to constraint which should be subject to an #interoperability obligation”

V nice @CNNum summary of what should qualify as a “systemic” platform in the #DigitalServicesAct! ?? (p.22 of my Google translation:…) Image

Social graph interoperability/portability “not considered sufficient by most of the interviewees. Indeed, even though the user who migrates to a new platform would keep his contacts, he still has to be able to continue contacting them.” (p.27)

“If the #interoperability of #InstantMessaging is generally perceived by interviewees as a necessary first step towards the interoperability of #SocialNetworks, it would not be sufficient, as it leaves out many basic features, namely everything that affects published content.”

“it emerges from the hearings that a platform must be able to protect its latest features to differentiate itself from competitors and attract users. The functions most likely to be made interoperable are therefore the most established” — eg those covered by W3C’s ActivityPub

The objections to “Interaction with content” (p28) are trivially overcome. Eg “technical difficulties” — use ActivityPub. “Conditions of use” — invalidate exclusion by dominant platforms. “each social network has specific interaction methods” — yes, leave out unique features

“the direct cost [to] platforms…of interoperability was judged to be low to moderate according to the different actors…
– development costs, which would be insignificant;
– day-to-day management costs, which may involve company adaptations in terms of training or recruitment”

Oh dear. #Interoperability is “excessively aggressive for the business model of large platforms” for @cedric_o. First time I’ve heard a European minister concerned for Zuckerberg’s $100bn wealth. What about the “value gap”, and “sucking sound” of transatlantic data flows? ?

The economic analysis is weak (p31). “several interviewees believe that interoperability could provide more data to large platforms, due to the importance of their social graphs” —> why strong #dataprotection rules are essential, and involvement of DPAs (like @CNIL)

“for some emerging social networks, interoperability with large platforms could be seen as a potential intrusion” << I don’t think anyone is proposing #interoperability be imposed on startups/small-medium-sized firms

The innovation analysis is even weaker (p31). It cites Furman, but misses most of that review’s points… (or the much more detailed @CMAgovUK analysis, which was surely available in draft?)

Important point: “according to @Conseil_Etat, the implementation of #interoperability could be reconciled with the freedom to conduct business, as long as it respects the principles of necessity and proportionality”, citing the CdE Opinion on the February @Senat bill

The IP issues are all easily dealt with (p32) using eg Standards-Essential Patent rules and consideration of responsibility between platforms for copyright or other infringements, and distributed takedown functionality

“an identity common to several networks does not prevent multiple identities. Users can, if they wish, continue to segment their uses & digital identities. For example, in the telecoms sector, interoperability in no way prevents to have several numbers (eg personal and business)”

Mmmm. I think just about every objection raised in the “uncertain balance” section is overstated and/or easily dealt with ?

I think the difficulties in using the European Electronic Communications Code are significantly overstated (having talked, recently, with some of the main authors of that code…)

.@CNNum: ” it would be better to examine, as a 1st step, the effects of the implementation of the right to data portability” << this should have happened already over two years ago as #GDPR came into force (Art.20), and anyway insufficient given network effects, as the report says! ?

I agree with most of the rest of the recommendations, except would go faster and further. The exhaustive evidence is now there to do this! Use the EECC to mandate messaging interoperability, and the #DSA to set in train immediate #interoperability mandate of ActivityPub ✌?