XFN – XHTML Friends Network (2003) | Hacker News Reader

XFN – XHTML Friends Network (2003)

gmpg.org

58 pointsthinkingemote15 comments

XFN is the origin of the rel=me attribute [1]. Rel-me adoption is growing in recent years, with Mastodon and others adopting it as a way to associate profile pages together [2].

Maybe two years ago, I added rel-me to a blogroll web crawler I'm building [3]. Unfortunately, theres a bunch of misuse, like people adding bidirectional rel-me to a company page which make it look like the whole company is one person.

I don't think we'll see other attributes, if only because people are more privacy aware. XFN is like using Facebook but letting every web crawler harvest your social graph, instead of just Facebook.

[1] https://developer.mozilla.org/en-US/docs/Web/HTML/Reference/...

[2] https://docs.joinmastodon.org/user/profile/#verification

[3] https://alexsci.com/rss-blogroll-network/discover/feed-de42f... (rel-me adds the checkmark here)

This brings to mind FOAF: https://en.wikipedia.org/wiki/FOAF

The early Web 2.0 era was still heavily riding the idea of the Semantic Web, which is why it saw the rise of a lot of metadata standards and "microformats". Most of these rapidly died when it became clear that there are stronger commercial incentives to abuse this data and falsify data (e.g. for "SEO", i.e. spam and false advertising) than there are social incentives to disclose this information and use it responsibly. It was a major clash between the idealistic vision of building global interconnected communities and the financial reality of companies wanting to optimize everything for their own profits.

In other words, the reason we can't have the Semantic Web is the reason social media has become largely unusable (cf. Dead Internet Theory) and the reason Facebook first decided to push features that they knew made their users measurably unhappier (the introduction of the "social stream" home screen instead of having to visit each connection's "wall" to see what they've been up to massively drove up engagement and created advertisement opportunities but resulted in users experiencing an expectation to "perform" for their "audience" instead of "sharing" with "friends" like they had done before - downstream this has resulted in mental health crises especially amongst teenagers using image-heavy social media like Instagram dominated by influencers).

I'd even go so far as to say that "social media" in its original sense died with the hopes of the Semantic Web and the end of the "blogosphere".

At the time it was called "social" media because there was a perception of building actual 1-to-1 connections between individuals (even prior to "real name policies") and people would share unadulterated personal experiences with the understanding that they were sharing them in a globally distributed but ultimately highly specific niche (and often using several such niches for different topics which is why tag clouds and such became popular). The shift away from this to "walled gardens" and "social feeds" (rather than curated RSS feeds and blogrolls) allowed for "brand accounts" (the most famous example likely being Wendy's on Twitter) and eventually the widespread use of Mechanical Turks (i.e. people in low-wage countries being paid to drive up "engagement" and steer perception via fake comments) and eventual full automation or "bots".

I guess FOAF stands out among others like XFN in that it's still seeing some use as a metadata standard because it defines vocabulary that remains useful outside its original intended purpose, surviving it.

The specs for things like microformats were mostly fine. The issue is that search engines mostly obfuscate their behavior and microformats fizzled out in the places that could have benefited from them. This was mainly due to relatively low adoption; not necessarily abuse. This was the kind of thing that techies obsessed about that most other people ignored.

So, search engines couldn't rely on sites properly using microformats on most sites and had to resort to other things to pull out structured information. Browsers removed what little support they had for microformats. And then blogging became more of a niche thing as people moved to social networks and things like substack, tumblr, and similar sites.

Mostly, you can still use microformats. It doesn't hurt anything and there might be some things out there that benefit from having a little bit of extra structure. But otherwise there's just very little point to it (beyond some niche uses); and there never really was. It's the classic chicken egg problem where it would be great if everybody used it but because most people didn't it wasn't all that great. You only get network effects if you get a critical mass of adoption. That just never happened with this stuff. Most developers treated it as a low value, optional thing, that was maybe nice to have but not worth obsessing over.

These days there are related things that are still widely used (e.g. opengraph and similar meta attributes). Also there are new semantic html tags that help tell apart main articles from navigation, footers, etc. that are fairly widely used. That helps things like screen readers, rendering previews of shared links, etc. And it also helps for SEO. And SEO is a great incentive for people to adopt stuff.

XFN and FOAF were indeed mostly relevant in the pre-social network era and became quickly irrelevant. The point with social networks is that they are walled gardens. Standardization was not something that the owners of those actively pursued because they are simply not interested in interoperability. Exclusive ownership of the relationship of the user is the whole point for exploiters of social networks like meta, tik tok, x/twitter, linkedin, etc.

> It was a major clash between the idealistic vision of building global interconnected communities and the financial reality of companies wanting to optimize everything for their own profits.

I think its even simpler than that. HTML et al was a nice idea but its largely ignored in the pursuit of making web pages seem like apps.

Berners Lee had this idea of documents like pages in a book, nothing like the full-fledged applications in the browser we get today, yet that legacy Berners Lee cruft still remains whilst people find ever more inventive ways to overcome it

I remember this being somewhat talked about in 2005. It was like an alternative to Friendster. But no one ever made a feature from this. Is this still a thing?

It’s a barely hashed together spec for associating webpages. It doesn’t seem to offer any of the “value” that social networks do, the prime of which being an easy-to-initialize profile with copious features.

It could be possible with nerds/geeks. However, that group (in the early 00s, at least) was far more interested in keeping their online anonymity prime and avoiding social networks; so I could totally see why it was dead in the water.

> It doesn’t seem to offer any of the “value” that social networks do, the prime of which being an easy-to-initialize profile with copious features.

It might be this is a feature, not a bug. Making things easy did help a lot of people connect in at the start, but connecting in the careless also has hazards, much less the content casino digital media has become.

Which goes into paragraph two, if you continued reading.

Around 2007 or so I worked on a semantic web search engine.

Among other things it ingested xfn (and foaf) so you could reasonably look for things like the connection between you and someone else who had a webpage, but the whole thing by itself didn't replace in any way what social networks were already offering, e.g. your feed, or direct messaging.

I think WordPress still supports xfn links, but it's not particularly useful.

>the whole thing by itself didn't replace in any way what social networks were already offering, e.g. your feed, or direct messaging.

It didn't get to that point, but there's a lot that could be possible if the graph were owned by everyone instead of just the social network owners. Distributed trust, moderation, etc.

Yeah that's always been the dream, but I'm afraid at this point I'm convinced noone wants that, because everyone just wants content to be consumed on their platform, not be available for everyone to reuse and remix on their own.

That's imo the real reason semweb, microformats etc.. failed. We're left with a bit of schema.org that just plays into Google's SEO tactics.

> what social networks were already offering, e.g. your feed, or direct messaging

by social networks i assume you mean specifically facebook, twitter etc and not IRC or any of the other pre-existing social networks that simply didnt have profile pages or comments sections

> Is this still a thing

Today's webpages are mostly a big pile of javascript generated Single-Page-Application™

So no it's not a thing. The HTML is barely parsable.

Dang, I was expecting a 1-900 phone number that would allow me to get my horoscope, sort of like the Psychic Friends Network from the 1990s, but in XML format.

Reply on news.ycombinator.com