Birb and Fossil: An RSS Revival? | Hacker News Reader

Birb and Fossil: An RSS Revival?

timkellogg.me

49 pointstkellogg54 comments

> notorious for sticking with chronologically-ordered timelines, so unless you have time to look at every single post, you’ll likely miss something.

The value of RSS to me is getting everything in chronological order, so I (not some algo) can throw 99% of it away, unread.

(and if I do want to go back and look at something I'd skipped without reading, it's easily findable, searchable by keyword or at least still there in the "read" list, next to all the things that were temporally close)

I don't understand that reasoning. If something is chronologically ordered I can scroll back to the last thing I saw and catch up, or likely not and just accept that I can't read everything, if it's interesting it'll probably be boosted or involved in ongoing discussions.

If it's ordered by algorithm I have no way of knowing what or how much I've missed.

Depends on the "algorithm"! At this point "algorithm" in the context of social media doesn't mean anything sensible. A very common "algorithm" used to order messages in email inboxes is by "read/unread" status, which very much helps you catch up efficiently. Chronological sorting is very limited, in comparison with the myriad of other ways you could organize such data.

I use newsboat, so it picks everything up in chronological (modulo feed oddity) order, but displays only "unread" items to me.

On top of that is further organisation: non-HN feeds are batched together in a single display, but HN often produces more than a screenful (~70) of unreads, so then I have hotkeys that filter out alphabetic subsets: A-G, H-R, or S-Z. That normally suffices to avoid any scrolling, but when I haven't been following RSS for several days, I also have filters that further limit to a single day, either 1, 2, 3, or 4 days ago.

Although the underlying data is kept chronologically indexed by newsboat, and I can always refer to it that way (frex the concrete splits AG/HR/SZ were determined by querying newsboat's SQLite instance) it by no means prevents layering a myriad (well 20, not counting newsboat's ad-hoc command line filter interface) of other ways to organise on top.

PS thank you for reminding me to finally look up how to turn off threading (View»Organise by Conversation) in the mac default mail proggy!

The best way is threading, which combines chronological, unread status, and the tree structure of replies.

Everything else is either an attempt to do less work or an attempt to shape user experiences for your own goals.

Threading requires two data elements per message, generated at the time of message creation: a unique-enough message ID, and the parent message ID. The client needs to have a method of storing 1 bit of state per message, either locally or on a home server. If messages have a timestamp, then a high-water mark is a useful heuristic. If the messages have required subjects, tags or topic groups, you can significantly improve both performance and user experience.

Usenet shall be resurrected, time and time again.

> At this point "algorithm" in the context of social media doesn't mean anything sensible.

Generally it means "select the material and the ordering in the manner that maximizes revenue for the platform's owners". I'd call that a sensible (if morally questionable) strategy for the owners.

I wasn't converted by algorthyms, but it's opening a new and exciting way to live my life.

For example, I've started ordering my calendar by algorithm. Whatever's at the top I'll go to, even if it was a week ago. Occasionally there'll still be some other people similarly sticking around in the concert hall or meeting room. Oh boy, do they have some stories!

I've also started paying bills by algorithm. If it's important enough, I'll pay it twice. Because algorithms at identifying the products I really need to pay for. If the algorithm doesn't prioritise it, I let it slide. I feel on the vanguard.

That's the only logical way. Give it a try.

There should be a "shuffle by site" option though. I tried using RSS and noticed that the feed is dominated by 1 or 2 sites that post daily, and it takes a lot of scrolling to find sites that post less often.

The whole "death of RSS" idea seems like a strange perspective. RSS/Atom never went away, and nothing has replaced their use case for syndicating content.

To the author's point, it's just that a lot of mass-market consumption of web content for the past decade or so has been on walled-garden platforms that never offered syndication in the first place.

But even there, I think the article is overstating things -- it implies that Reddit, HN, Medium, and Substack have only recently begun offering RSS feeds, but these have never not offered RSS feeds (and HN has always had a native top-level article feed, even if the 3rd-party solution is more extensive). Even YouTube has always offered RSS feeds (albeit without enclosures, so they can't be used as podcast feeds -- but Odysee, a web frontend to LBRY which is gaining traction against YouTube does offer feeds with enclosures).

I guess this application is a good solution for people who want to follow syndicated content via Mastodon, but it should be pointed out that the traditional model of using standalone RSS readers never went away -- when Google Reader was shut down, the void was quickly filled with a variety of solutions like TinyTinyRSS, MiniFlux, Feedly, Inoreader, etc.

I personally use TinyTinyRSS, with Liferea as a desktop client, as my primary interface to all of the blogs, podcasts, subreddits, and YouTube channels I read, along with aggregators like HN and Lobsters. I've been using this solution for over 10 years now; nothing has ever stopped working and none of the sites ever pulled back from publishing feeds.

Just as an FYI, the Odysee frontend might be on the way out unfortunately [1]

[1]https://www.theguardian.com/technology/2023/jul/16/lbry-clos...

Looks like FUD. The original LBRY shut down, but Odysee was spun off into its own business entity, which appears to be very much in business, and the current home page doesn't seem to contain a trace of "far-right" content. The underlying platform is a decentralized blockchain, and will maintain continuity regardless of what business entity is acting as its front face at the moment.

I always thought, the "correct" way of doing this was the other way around: the RSS reader would implement ActivityPub so you could "toot" from within you RSS client. It would perhaps attempt to collect other "toot"s about the same link, facilitating a discussion (and keeping everything at one place). But I think this is the next best thing, especially for those feeds you tend to share on the social networks (I don't think it's feasible to reproduce the RSS reader following with this).

That's kinda how we've approached it with some of the IndieWeb tools (https://indieweb.org/reader) where you can subscribe to the various feeds you want to, and can directly like/reply/etc

I am an active user of Monocle (https://monocle.p3k.io), a social reader UI (which is connected to a feed polling service, in my case Aperture). I can click "like" or "reply" and the reaction is published on my personal website. A notification is sent to the recipient via Webmention that I liked/commented on their content (if their site supports webmention).

> It would perhaps attempt to collect other "toot"s about the same link, facilitating a discussion (and keeping everything at one place).

If you go to Explore → News on Mastodon, there's like a list of 10 trending articles within your network. It also includes "discussed by X people in the past Y days", but annoyingly there's no way to reach those discussions and read what people thought about the article.

So I feel like Mastodon's moving in the right path to serve that purpose, it's just unfinished.

Isn't that effectively Micropub/Microsub? RSS was only ever meant for passive consumption, but similar IndieWeb specs have been layered on to support more social features and two-way engagement.

Many RSS readers implement a "share" button that can be used to post to a mastodon account. Is this different than the workflow you're describing?

Yes, there is a slight difference in that, I would like the RSS reader to find other social accounts (Mastodon, Lemmy, etc.) that shared the same link and bring them in context so that the user can directly interact with those as well. I know it's not trivial to do, but one gets to dream, right?

Stop trying to make Mastodon be Twitter. If that's what you want, go use Twitter. I don't get this mentality of "I like X, but I don't want to use X, so imma go turn Y into X" some devs seem to have. Chronological ordering is a fair and balanced choice for media presentation. Playing into people's FOMO is what got the Internet into the rat's nest of algorithmically driven problematic feeds in the first place.

eludes me how people think it's ok to tell other people to stop on topics like this. it's not even about making maston be twitter. it's about experimenting with timeline options that are not chronological and it's all open and an invitation for people to contribute if they are interested. you are not interested? cool, then do whatever. because unlike twitter, you have that option - and that option will not go away.

stop commenting when your comment tells others to stop doing what they like because you don't like it.

Oh, grow up. Dissenting opinions are welcome on line, being probably the most beautiful part of this whole mess we call the Internet. If you don't like my comment, address the content of the comment and explain to me how my warning of the path this treads is unreasonable. Stop virtue signalling and contribute. I'm open to having my mind change and everyone on this site should be as well.

dissenting opinions are not a problem until they start telling people what to do.

i'm not interested in changing your mind.

Interesting. So you wanting me to stop explaining why someone shouldn't be doing something is you not being interested in changing my mind? You live in an interesting world.

What is a dissenting opinion other than an expression of dissatisfaction toward a thing, be that an action of another or not? You say dissent is okay, byt in the same breath say byt only if you do it the way I think it should be done.

Hypocrisy is a helluva drug, I guess.

Anyway, have a great day. Didn't mean to piss in anyone's cereal quite this early.

everyone can dissent as much as he likes and make it public too. personally i'd have preferred if it wasn't that imperative. it's not what, but how one says things. but anyway, maybe that other commenter is right that it shouldn't be seen that emotional.

Using the imperative form in writing is more commonly used as rethorical device than as a literal order. You interpreting it as the latter suggests to me that maybe you should work on your emotional regulation when it comes to other people's writing.

Dissenting opinions are welcome with me, I appreciate them. What I don't appreciate is casual downvoting. I suspect I must have some of the highest downvote stats on this site. And I would love to be able to search HN users using that sort of metric!

Anyway, I have unpopular, but sincerely held opinions - anyone is free to disagree, of course. The point I'm trying to get to is that while I hardly ever downvote, if I did, I would hope that I would also comment on what the issue was.

As a courtesy, it should be downvote + comment, if poss.

I don't really know how you migrated to the subject of downvoting, but I do not disagree. Never found the idea of "karma" on a forum to be useful for anything other that brigading and collective censorship. I use Harmonic to view HN, which has options yo hide scores and all that, do I don't even see the downvotes, tbh. Safe to assume they are there though, especially when you say things that do not jive with the popular narrative.

have an upvote!

I think some subset of people do want an algorithmic Twitter-like thing which is not centrally controlled by a crazy person. I know people who prefer Bluesky to Mastodon for precisely that reason (though Bluesky's lack of central control feels very... coming-soon, rather than actually real).

>Stop trying to make Mastodon be Twitter. If that's what you want, go use Twitter.

I don't see how using an optional, open-source, non-chronological feed means that someone should just "go use Twitter". For one, the incentives are completely different between Twitter which generates revenue based on engaging users (and juicing subsequent algorithms), and someone who builds a separate feed interface for Mastodon.

>I don't get this mentality of "I like X, but I don't want to use X, so imma go turn Y into X" some devs seem to have.

Developers like to solve problems and build things. I've re-created other projects and products lots of times. Why is it a problem what someone chooses to do with their own time for their own reasons?

>Playing into people's FOMO is what got the Internet into the rat's nest

Disagree. Incentives tied to advertising-based business models is the root cause here, not FOMO.

Disclaimer: I made my own alternative timeline for Mastodon as well, so I'm pretty biased. :)

Bias aside, you present fair counters to my position. However, I still think the leveraging of FOMO is what drives us in the wrong direction with social media. If the market (us, basically) had the general attitude of "if I miss a post, I miss a post, oh well" the advert-based models wouldn't be as effective as they are, incentive would shrink, and we'd probably have an Internet that was closer to the original intention. That, however, is strong speculation on my part and I am hyper-aware of that, but having watched social media (as an adult) emerge and become what it is today, it's exceedingly difficult for me to unsee that pattern.

We can perhaps agree to disagree on that, but I'm considering your other points.

Just my two cents, but I've always viewed FOMO and the ad-based algorithm as two sides of a feedback loop.

FOMO is the psychological response that the ad-based algorithm is playing off of, neither one is the problem in and of itself.

> I don't get this mentality of "I like X, but I don't want to use X, so imma go turn Y into X"

You really don't understand why someone would prefer Mastodon because of its decentralized and open source nature, but would like to take some of the things they like about Twitter and include it in their Mastodon experience?

> Playing into people's FOMO

This implies that the only form of "algorithmic feed" is one where clickbait and inflammatory content is maximized.

An algorithmic feed can also mean something like a "for you" feed that tries to find you posts you'd find interesting based solely on your current follows, but from other accounts you're not following.

It can also mean a "what you missed" feed like the one Twitter used to have, that shows you the top posts from your follows since the last time you logged in. If someone doesn't open mastodon for a week because they were on vacation, it's not feasible for them to scroll through an entire weeks worth of posts to "catch up".

Understand that people have a wide variety of experiences they'd like to get out of social media, and that this whole mentality of "my idea of a social network is the most enlightened and correct one, and everyone else is stupid" just feeds into the stereotype of the socially inept engineer.

Totally get what you're saying and agree to an extent, but can you supply examples where algo-driven feeds did not eventually devolve into clickbait/inflammatory/etc recommendations?

I am not saying algos are bad, so I should clarify that. What I should have directed my ire at is the people that implement them since, historically speaking, they can not be trusted to do so without bias or other motives. Chronological feeds cannot be abused in this way, to my knowledge, and represent a state of purity in my mind.

Appreciate you taking the time to make salient points, though.

The alignment problem is an extremely difficult one, and one that we've basically punted on for all AI research. Making an algorithm is easy, scoring content by external factors like engagement or time on screen are also easy. Designing a scoring algorithm that is aligned to a person's preferences, desires, opinions, etc are extremely hard. At best we attempt to find proxies and correlations, the algorithm tracks engagement and we assume that means they want to see more similar content.

IMO this is the fundamental reason for chronological feeds. Aligning the algorithm is an unsolved problem, I'd rather just not use it at all.

Twitter's algorithm optimizes for advertising. I want my algorithm to optimize for my value.

Use case: I want to log into Mastodon once or twice a week. I want to follow lots of interesting people and topics, but I don't have time to read it all. So, I want to have an algorithm that allows me to see all things from a subset of accounts I consider friends and, after that, see the most shared and starred from the larger group I follow, then the most shared and starred from the topics I follow.

But if I happen to be browsing more frequently, I can see the chronological feed again in that order: close friends, acquaintances, topics. This isn't what Twitter does: their algorithm does not serve me.

Good counter, you and a few others have given me something to consider.

Nothing about this mimics Twitter. What the heck do you care what somebody does with open source software in a part of a federated ecosystem that you don't have to use? The whole point of software is to serve people's needs; don't tell people what their needs should be. If some person wants to have a non-chronological timeline, and others want to use that same software, that's their prerogative. If that makes you mad, there are plenty of trivial ways to wall yourself off from all of this.

Perhaps you should have walled yourself off from my comment, but you still chose to get upset and reply. Weird how that works.

I'll you the same thing I told the other responder; Hypocrisy's a helluva drug, eh?

Anyway, moving on.

For a moment there I thought you were the first person to actually follow the new branding guide for the platform formerly known as Twitter. Of course, I should've known better.

Name clash with an already popular open-source project: https://fossil-scm.org/home/doc/trunk/www/index.wiki

Which is also a name clash on plan9's fossil archival file server/system

Both these projects are fossils at this point.

/joke

ha! I just posted a comment saying that very thing...I was confused.

Me too - I expected hosting an RSS feed via Fossil or something.

I wish people did a search <project name> open source before naming their project

People like RSS because it's not algorithmic. People like Mastodon because it's not algorithmic.

What I'm mainly curious about here is what drew this author to these two technologies in the first place? What's the hook for them?

---

Subjective aside: before I read the line on algorithms & the HN comments bemoaning the same, I was already very turned off by the awful AI header image. Until tools like Dall-E, &c. get to the point where they can reliably generate images that aren't blatantly & obviously AI-generated (they seem a surprisingly long way off still), I think this effect is worth keeping in mind: for me the aesthetic is a massive turnoff for any product page.

Some people like Mastodon because it's not algorithmic, some people like it because it's not centralised/has no real single point of failure, some for other reasons. I'd definitely buy that at least some people would quite like an algorithmic view of their Mastodon feed; wouldn't be for me, but one reason that people go for Bluesky over Mastodon is that it is algorithmic.

> one reason that people go for Bluesky over Mastodon is that it is algorithmic

My comment does seem a little dogmatic whereas there's always a broad variety of approaches & preferences with anything. I was somewhat exaggerating, but my goal was to represent the typical user as best as I can glean.

As someone who advocated heavily for Mastodon (& continues to), I spend more time on bsky now, but I've found people's reasons for being there over Mastodon to be either my reason: network effect, or for the vast majority of folk: UX. ActivityPub client/server UX remains abysmal in comparison to centralised services (not complaining as I know it's a harder problem for decentralised tech, just stating the fact).

Algorithmic feeds haven't come up in the context of bsky prefs mainly because - in my experience - algorithmic feeds in any app/website are broadly hated by all but the most tech-savvy of people.

Actually, I wouldn't mind a little bit of algorithm in my feed.

"Newest first" is annoying because unless I manage to read everything that's new then I have gaps. If I've only got a couple of minutes now then I want to be able to see a few things without losing track of what I have and haven't read.

"Oldest first" is annoying because -- as much as I might want to -- I can't keep up.

"Newest unread" is annoying because if someone's posted a series then I want to read them in order.

"Prioritised unread" might be OK -- if someone only posts once a month then I probably want to prioritise reading their post over reading absolutely everything by someone who posts multiple times a day.

I'm currently reading my RSS with Newsblur, and it's not perfect but it does give me flexibility through the "intelligence trainer" to highlight the stuff I want to prioritise.

Aside, RSS and Mastodon definitely are algorithmic. "Chronological ordering" is an algorithm, and it's not always the most appropriate.

Initially I thought they were referring to fossil SCM:

https://www2.fossil-scm.org/home/doc/trunk/www/index.wiki

Since fossil SCM has been around awhile, the author might consider a name change to avoid confusion?

Fossil watches has existed since the 80s. The Fossil wrist PDA came out three years before Fossil SCM was first released. Why does Fossil SCM get to lay claim to this name?

Ten seconds of reading and you'd have avoided any confusion.

started using RSS feeds a few months ago, they're still widely supported and there are JSON scrapers for those sites that don't. The idea of "reading N accounts" to get some info or some algorithmic soup of "popular stuff i tangentially show interest in" doesn't even compare to curated RSS feeds organized into groups. The unpopular, rare and obscure stuff can't compete with algorithmically optimal SEO-friendly hype-tagged social media junk in the "common prole-feed" of giant websites.