Greenwich: an experiment in collaborative links
readpolymathematics.substack.com
readpolymathematics.substack.com
One suggestion is to maybe allow users/community to have a walled garden in regards to writing rights. This would help with moderation, and allow users to subscribe to the walled gardens/bubbles they are most interested in.
https://www.w3.org/TR/annotation-model/
It deserves more attention, particularly from browser vendors and social media platforms, but the incentives have never been in place.
One commercial application built on the standard is hypothes.is, but I’ve lost track of their efforts years ago.
Recently, I came across this paper which describes an implementation called "Weblinks" (terrible name) that focuses on just the links. Haven't used it, but it's thoughtfully designed:
Once it can be on Firefox at least, I can try it.
(The example that came to mind first was Google Sidewiki, but it looks like there's a bunch of these listed in https://en.wikipedia.org/wiki/Web_annotation)
But yes, it seems a bit like a small twist on an extremely old idea.
I commend that to the author in the spirit of learning about the space and thinking through the implications, because I believe people are more likely to solve problems when they understand what they are, and think through them, and don't just try to blunder past them with hope and moxie.
That post was written about generalized text annotation and I stand by it in that context.
However, as you deviate from the system being analyzed, the analysis becomes less appropriate. One of the problems generalized annotation systems have is that there are an arbitrary amount of textual comments that can be added to a page. That is what turns the popular pages into a unpleasant cacophony. The range of links is somewhat more restrained. Plus, links are just... links. Textual comments are arguing and flames and generally tiring on any popular page. (Though I'd watch out for people learning how to turn "links" into arguments.)
It is possible that a shared cross-linking system might work, but I'd strenuously suggest thinking very very hard before adding in any sort of inline "conversation" system. It is very, very obvious and very, very tempting... and it immediately puts you back into the generalized web annotation space, which is strewn with corpses, many of them very very well funded. You can have a "community forum" where people can talk, and perhaps even should, but putting it inline on the page is basically a known-fail. If you think you've got a solution to that I'd try to be very sure that you've got a very strong proposition on exactly how yours is different than the previous attempts.
Anyhow, I would just generally suggest to the author that this is one of those "obvious ideas" that hasn't happened because in general they flame out so quickly that you don't even find out they existed before they've already collapsed, not because nobody has ever tried it. Be sure to consider what has happened in the past.
Also, as a hint from previous efforts: It looks like you may be trying to bind links to specific text on the page. As you've probably already discovered, that's harder than it looks, and however much code you've written for it, I guarantee you it's even harder than that. I'd strongly suggest considering just binding links to pages and not trying to bind it to text at all.
So I guess “clickbaity title” maybe actually means “clear vision to attract people” and “good storytelling to engage users”.
I remember there was a website that did this in 1999, using frames to allow people to post comments on any website. The courts shot this down as an illegal infringement of trademark. Does anyone remember the name of that website that did this?
The potential for abuse here is enormous. I have a difficult time seeing this becoming anything other than a cesspool of ads, 4chan-style joke links, and general inanity.
Like I don’t trust the internet in general to curate these links. But one could surely find networks where the average voter could be trustworthy…
You can look at the full user tree (https://lobste.rs/users) and see if there's anyone you know who might be willing to invite you to join.
I have been a good citizen of Hacker News for fourteen years now...
Anyone here would want to throw me an invite to lobste.rs?
:)
At least for me, I'm the only HN user I know except my dad who doesn't even post, he just got lurk links from his knee if the woods like hackaday.
Modern problems...
It's basically moderation, or a subdomain if you will, and I'm not sure there any place or product that currently has fully automated moderation that works. There's always a human involved if you want to do it properly.
Moderation of, say, a comment section has some problems that seem (to me at least, I don’t actually work in this area, so maybe this is a naive take) to make it much harder.
* There’s a desire to preserve continuity of conversation
* People have different expectations of the types of content they find to be outside the domain of reasonableness
* People have different expectations of what the job of moderation is (curating productive discussion or just banning truly odious stuff?)
Like if I say I’m going to only host technical discussions, we will get a spiraling argument about where exactly the line is between tech and political policy around tech.
The easy solution is for users to just mute people that they don’t like, but them you have a conversation where some participants can’t see eachother, they managing back-and-forth a between people with different muted subsets.
And there’s still the issue of managing the general vibe, if it becomes conventional to throw around unpleasant or hyperbolic language, that could ruin the discussion for everyone, even those who’ve blocked the main perpetrators. Or the vibe could become toxic to new users who haven’t curated a blocklist yet.
In this case, there’s no need to preserve the continuity of conversation. And the users have less ability to continuously change the vibe, since it is just a collection of links. And the entry point could be trusting a single user, so the overall vibe is less relevant.
There are studies showing that comments on articles erode the trust readers have in these articles. Given the quality of the average comment, it's likely that comment systems on most sites make people both dumber and angrier.
So I think the idea of forcing comments (or user-contributed links) on sites that don't want comments is fundamentally problematic.
Personally, I don't want random people on the Internet putting links on my articles. If you want to discuss what I write, or provider additional context, or disagree, then do it on your own site, or in a public place like Hacker News, not on my site.
How many users will take the thing for a test drive and then patiently click through and downvote every goatse link they find versus simply uninstalling it when they get goatse'd?
IMO this is the kind of content which made early 2000 internet fun. Not the bland, moderated to hell and back social media sites are. Just compare what happened with the million checkbox experience which got a secret ARG made by users having fun and reddit place which is... meh.
However, would it be possible to come up with a UI that doesn't need to "Read and change all your data on all websites." ?
I feel uncomfortable with that.
What is the plan for fighting bitrot and bad-faith actors?
This turns it into a centralized problem, but a problem nevertheless?
I would think scammers and spammers would just focus on lucrative targets like instafacetok and yelpazon reviews in 2024. And is seo hacking based on comment threads really still a viable business, or we are mainly worried about links to malware?
I get that new platforms of any kind are certainly likely to be targets of vandalism, but it’s surprising that everyone seems to be suggesting that every new/unpopular/niche platform will immediately be targeted by what amounts to organized crime. Even if I had evil intentions and an army of thousands to craft malware and scams, I wouldn’t task even one of them to poke around on platforms with less than like 10% of the market, because why bother?
Yes, the most basic ones just go by form fields, easy to see if you expose some sort of comment form and add a hidden "first_name" field users don't see, but the bots are stupid enough to fill out all fields no matter their visibility.
The more advanced ones won't get caught in that, but there are tons of general tools for automatically spamming even hand made forms/fields on the fly.
I can think of a few reasons.
I can see how webrings died when social media took over — not that I believe social media to be superior in any way — but I never understood why making webpages editable never took off. There were a few attempts, most requiring extensions.
EDIT: Chrome-only extension? Now I don't feel so bad about those thrown word-turds ;)
Also, it'd be great if this was an open standard so that anyone could write an extension for any browser.
Ignore these insufferable know-it-all haters on Hacker News. These people are the worst! Some aren't even real people.
The ones that are bots are probably HN itself because https://greenwich-for-chrome.replit.app/ is a threat. This is the coolest part IMO, has Twitter or HN like potential.