More Words on Webmentions (and Backfeed)
Webmentions arenât actually the problem
After my earlier post on webmentions, Cam kindly reminded me that the bit I felt uneasy about is actually backfeed. To quote the IndieWeb wiki:
Backfeed is the process of syndicating interactions on your POSSE copies back (AKA reverse syndicating) to your original posts.
Itâs interesting reading the discussion section on the wiki page, as it acknowledges some of the concerns that have come up of late, but ultimately, skims over them.
The list of âproâ reasons for backfeed mentions making it easier for people to reply, particularly those who are less technical. However true this might be, I believe itâs probably equally true that the less technical someone is the then less likely they will understand why their words are appearing on a different site. Their expectations[1] are almost certainly âI replied/liked/whatever on Site A (more likely App A), so itâs not going to appear on Site Bâ. Is it right weâre[2] subverting these expectations in an effort to make blogs and personal websites âmore socialâ?
Aside from the privacy and implementation concerns noted in the last post, Cam noted in his own post that another problem backfeed can lead to is context collapse. This isnât a problem I have much experience with (so canât really talk to it) but I thought it worth noting here for completeness.
This is not a new discussion
Terrence wrote about the ethics of this topic in 2022. His concerns were/are pretty much identical to my own. Bix wrote similarly on the integration of ActivityPub into everything just last month. Sebastian wrote an in-depth look at the problem - particularly through the lens of GDPR - back in 2018! There will be many more posts out there about the issues - and no doubt, potential solutions - to the problems with backfeed as a concept. If you know of any good ones, please send it my way on Mastodon.
Itâs on us as webmasters
(Yeah, I cited the deep magic. Substitute for âsite adminâ or âsite ownerâ if you prefer.)
How we handle the issues that have been raised is ultimately on us: the people who build and run the websites who integrate these features. We have to be more thoughtful about these things rather than rushing to âoh, cool new technologyâ. I didnât really pay the issues much heed when I first implemented backfeed in 2019, but I should have.
There might come a day when users expect their details, words, and actions could show up on other sites they werenât aware of. ActivityPub and other forms of the Fediverse and federation might herald that future. But I donât think weâre there yet, and Iâm not convinced everyone will buy into it. Iâve seen enough threads on Bluesky where someone reacts with abject horror when they learn that basically everything on Bluesky is public, and similar threads when someone learns just how far public posts on Mastodon can reach.
By all means send and receive webmentions. Even backfeed interactions onto your site. Just be mindful of how you do it and whether displaying all the data you receive is the best thing to do.
A quick clarification: I â¤ď¸ Brid.gy
I wrote my last post pretty quickly and late at night, immediately before going to bed. After a re-read this morning I felt that it occasionally came across as me having a go at Brid.gy. That wasnât my intent; I admire the technical feat of what Brid.gy achieves and could achieve in the future, and I respect that it is an open service helping to link disparate parts of the internet together, for free. My issues really lie with the consumption of what Brid.gy provides, rather than Brid.gy itself.