Should W3C Social Web Interest Community Group (open to all) resolve next week to charter a new Social Web Working Group (open only to employees of member companies) without clarifying why or to what ends?
Context:
It was almost proposed/passed without notice at unrecorded and hard to hear remotely in person meeting the other day at TPAC.
https://mastodon.social/@bengo/111051525999167046
It will be proposed on Sep 22:
https://lists.w3.org/Archives/Public/public-swicg/2023Sep/0012.html
Boost it
@bengo "It was almost proposed/passed without notice at unrecorded and hard to hear remotely in person meeting the other day at TPAC."
That's not true.
The proposal was to start the chartering process. The CG can't set up a WG; we have to request one from the @w3c .
So, the process would be making a bullet-point list of what the WG would be working on.
@bengo @evan @w3c I read the minutes and can confirm that this wasnt really a close call. but im not sure i understand the benefits of having a separate (and potentially closed-off) working group in addition to the open CG, and I did not feel that the proposers clearly articulated their motivations behind this proposal. My gut is telling me that something's off.
Are there any valid arguments in favor of doing this?
@djsf @bengo @w3c the main one is that only a working group can publish updates to the official documents. The CG can publish errata and clarifications, but it takes a WG to make an official update.
(Mostly. Apparently there's a process where we can correct editorial errors, but not make normative changes.)
@bengo @evan @w3c that is a good first line of defense. generally i would think it is not useful create a WG unless it is trying to solve a widely recognized problem in the social web or with a specific document.
"Widely recognized" means that it has been a topic of organic discussion in the wider W3 community and not an artificial one invented solely for the purpose of forming a WG.
Confusion around the purpose of a proposed WG is not a great sign.
@djsf @bengo the widely recognized problem is that we have some assumptions built into the specs that aren't 100% clear.
For example, we don't say that the people in your followers list are unique; they should only appear on that list once.
It would be nice to add that kind of clarification to the documents. Technically, it's a normative change, so we need a WG.
I think you may be overestimating the interest of tech companies in ActivityPub.
@evan @bengo that makes sense and I can see how a WG could be appropriate here.
However I don't think it's possible to overestimate the interest that entrenched corporations have in protecting their bottom line. EEE is still a real threat to the Fediverse and I think we need to speak plainly about this and stay vigilant.
@djsf @evan @bengo Why are corpos even at the table? It should be instances and open source implementations that work on this stuff. Like what Facebook is going to come in and be like "Oh we need such and such thing in the standard" and the folks at gaycommunists.gunclub need to suddenly break everything in order to support it?
@evan @djsf @bengo Hmm I think maybe the difference is I'd like them to implement but not be in charge. The power dynamics would be so imbalanced compared to anything else. IMO it'd be nice if AP was more like HTTP where we had the core more set in stone unlike how stuff like Group Ware and email got turned into something non–corpo impls couldn't coexist with easily.
@evan @mauve @djsf @bengo "Lots of implementations" is not enough to prevent domination. If one implementation represents 90%+ of the AP Fediverse (like Mastodon does currently), domination is a real and significant risk. If Threads implements some variant of AP, they could potentially become the new dominant implementation.
@steve @evan @djsf @bengo Like when they inevitably try to add ads to AP or whatever shit we can give them the finger and fork from any impl that's shitty enough to go with it. Prople thay want big social media corps can go ahead and use em, and existing AP communiyies can do fine without it.
I think it's mostly the monolith instances like .social that will cave to incorporate whatever facebook wants of them
@steve @mauve @bengo hard agree. I think @evan maybe meant to say "a wide variety of implementations", which I would agree with. But corporations will probably create just one implementation and throw lots of money behind it in hopes of dominating the market (e.g. Threads). That will divert newcomers away from open source implementations of AP and likely have a cooling effect on innovation in the Fediverse as a whole.
@djsf @mauve @bengo @evan It's the same risk with a wide variety of implementations if one implementation represents a huge majority of the deployed servers. We have a wide variety of implementations now, but the Mastodon implementation dominates and defines the de facto AP usage for the Fediverse (at least for microblogging and status publication, which is the most common use case).
@djsf @mauve @bengo @evan Re: "throwing lots of money behind it"... if Meta gets serious about AP, they'd probably forego investing much money and effort into influencing the W3C and the recharted WG and instead try to add Eugen to the payroll (while still leading Mastodon development as a Meta employee or contractor). The typical annual compensation for a Meta principal engineer is more than Eugen can make in a decade or two with Mastodon. "Make him an offer he can't refuse." ;-)
@evan @mauve @bengo more widespread is great. but if Threads attracts enough users to become the dominant player in the Fediverse, they could start closing it off or changing it in favor of themselves and to the detriment of others. So i'd like to see a diverse array of small to medium sized orgs and companies or communities embracing AP instead. the involvement of large corporations makes me uneasy.
@mauve @djsf @bengo I think that makes a lot of sense, too.
I also think having it more widespread, with lots of implementations, makes it less likely that any one can dominate the network.