Re: Interoperability and competition [was: Google, etc, and their proprietary protocols]

[Mark Nottingham <[e-mail]>]

> On 1 Dec 2020, at 3:36 am, Vittorio Bertola <[e-mail]> wrote:
> 
> 
> 
>> Il 28/11/2020 02:52 Mark Nottingham <[e-mail]> ha scritto:
>> 
>> So, what's interesting to me here is that we have a very good 
>> possibility of at least one, and perhaps many, jurisdictions 
>> imposing interoperability requirements on various digital platforms
>> as a remedy for competition issues.*
> 
> I am quite involved in the European one, so I am happy to see this mentioned here. Also, it is rare to have opportunities to discuss this topic with non-Europeans and with people from the global platforms, and it is a pity.

It seems like there's been a lot more interest in places like Australia and even the US with recent reports. :)


>> A third option that hasn't been discussed as much is leveraging 
>> existing SDOs to do the technical work that enables interoperability
>> -- ideally using already-existing protocols and formats where 
>> necessary. 
> 
> This is the likely model in Europe, where the Commission already has procedures to "recognize" standards developed by SDOs and bless them. There would still be a regulatory authority (a single European one or a constellation of national ones) that would define the policy objectives, but then the SDOs would be scouted to find suitable existing standards, or to see if any of them wants to develop one.
> 
>> There are lots of open questions about how this would work, but 
>> first I wonder how the IETF would react -- if a competition 
>> authority came to us and said "if you create a standard that 
>> promotes interoperability for scope _foo_ in timeframe _bar_ using
>> your normal process, we will leverage that in our remedy", would we 
>> rise to the occasion? 
> 
> I'd also like to know. However, I'd say that if Europe ever wanted to do something like this, perhaps the first door that would be knocked upon would be ETSI, not the IETF. This is due not just to the basic fact that the "E" in ETSI stands for European, but also to the fact that, in the general perception, the IETF does not seem to like cooperating with governments, or even to recognize them a role.

That's interesting. I'd characterise it slightly differently -- the IETF does not like cooperating with governments on projects that are perceived as harming the Internet. From what I've seen there's pretty strong concern held by many here about consolidation of power on the Internet, and a fair amount of experience in areas that could be helpful (e.g., JSON formats, Atom publishing, HTTP APIs), so it might be a good fit.

ETSI is an attractive venue (that unfortunately I haven't participated in yet), but I'd be concerned that the membership rules and weighted voting procedures could affect the result; many of the potential work items here (e.g., APIs for social networking) would benefit from broad participation and a focus on end user needs, rather than being unduly influenced by large corporate concerns. I can imagine that some big vendors would still be attracted to unwieldy solutions like SOAP here...

Of course, if the work items were just lightweight profiles of existing specifications (e.g., Atom, ActivityStreams, etc.), that might be workable. I see that ETSI has the ability to enter into a Partnership Project; for some items, perhaps it might be worth considering use of that mechanism to involve orgs like the IETF.


> Also, there is indeed some concern that open, consensus-based SDOs where the global platforms are strongly represented could fail in producing any suitable standard, as such a standard would negatively affect the interests of a significant part of the participants, which would thus prevent consensus from ever being reached.  This explains, for example, why the bylaws of the GAIA-X Foundation(*) exclude non-European companies (including European subsidiaries of non-European groups) from applying for leadership positions.

Absolutely; that's a risk that's going to be very front-of-mind for venues like the W3C. However, if Europe tries to keep too tight of a hold on the reins, it's going to make it look like a regional solution, and increase the chances of fragmentation.

It's also interesting to me when the IETF is portrayed as being vendor-driven/owned. While it's certainly dominated by tech people, and they have a number of limitations, I think most vendors would say that they do not feel in control of the discussions here; there are too many individuals who prioritise the Internet overall over their employer at the moment (acknowledging that some of them prioritise their view of the technology, rather than the users of the Internet in doing so).

In any of these fora, the practicalities of decision-making under these conditions needs some consideration. For example, if a competition authority tells BigPlatformCorp that it must expose an API to be standardised by the FooWG, how does the FooWG consider input from BigPlatformCorp participants? What if those participants say that BigPlatformCorp can't implement a proposal -- how do you evaluate and weigh that objection (e.g., how much should their costs be considered)? Will appeals from them be handled in a special way? If the normal process isn't adequate to achieve the outcome you want, would adopting special rules to accommodate a mandate from a competition authority cause anti-trust issues in other jurisdictions? And so on.


>> Past history like DNT in the W3C suggests that there's a will in at 
>> least some standards fora to provide affordances for the law. 
> 
> I would not mention DNT to governments as an example of how this cooperation can work, given its total failure...

:)  Yes, that wasn't done well; it always seemed like an instance of hope-based standardisation. I think we'd all benefit if there was some more thought (and research?) about why it did, and how we could do better.

Cheers,



--
Mark Nottingham   https://www.mnot.net/