RSS Server Side Reader | Hacker News Reader

RSS Server Side Reader

matklad.github.io

46 pointsBogdanp51 comments

I love that we're rediscovering server side rendering. It's impressive to me how quickly the industry moved away from it as the default and is now an interesting, new technique to many.

Except this has nothing to do with Server side rendering.

What makes you say this? A server is rendering/outputting HTML as output from a GitHub action that builds via Deno.

Did you think your browser was assembling the HTML?

By that logic, a static site generator is "Server Side Rendering".

Rendered once, yeah. Not sure why you’re splitting hairs when the client still isn’t doing the work.

If I were really splitting hairs, I'd say that the "rendering" (i.e, transforming the encoded data into its visual representation) can only happen in the client.

But anyway, regardless of what you want to call it: this whole thing is just a silly exercise to get a bunch of XML and transform into (X)HTML. If it works for them... ¯_(ツ)_/¯

I read the entire article and I still have no idea what author was trying to achieve.

Instead of a regular RSS reader he wanted to only get a notification when some target sites have new entries. He opted to implement an expanded blogroll, one that also shows the last three entries.

I feel like per site notifications, or muting low priority feeds, should be table stakes?

I did pretty much the same thing last weekend. I have a small Common Lisp program that generates my website. Last weekend, I wrote a new program to generate a blogroll page from a list of feed URLs. You can see the code and the resulting blogroll here:

https://github.com/susam/susam.net/blob/main/roll.lisp

https://susam.net/roll.html

This program was directly inspired by @matklad's idea of using a blogroll as an RSS reader. It's only been a few days, but I already feel like I can finally stay on top of my feeds!

Yup. I'm no programmer by any stretch, but I wrote my own crappy little thing years ago and I still use it nearly daily. One thing I realized is that everything else I used had too many features; one I really never ever wanted was a count of what I had and hadn't read. I'd rather use RSS really as "maybe some days I pick up the newspaper and some I don't"

https://gitlab.com/jrm4/mahrss (I don't even remember if this is the last version and make no guarantees about any kind of functionality, but you can get an idea)

HN seems to be having trouble understanding why this is brilliant.

Ignore the parts about JSON/XML. That's irrelevant.

Problem: you want an RSS reader, but RSS readers are annoying because they are stateful and you have to try to sync them across devices. Or, as in the case of Google Reader they may be discontinued. Best case, you have a dependency on a third party application.

Solution: make a web page on your personal site that aggregates links from your RSS feeds.

This is handy because you can now simply access your own web site as an RSS reader. As a side benefit, you can share this page with your friends to help them find nice links, and help promote stuff that you like to search engines.

Problem: you want to read your RSS feeds, but you aren't online.

Solution: use something like The Old Reader, which aggregates online, and can also be synced with an app like GReader for offline reading.

When are you ever offline these days though?

When travelling by air it's nice to have a bunch of feeds to read.

I have found significant issues with this. Many websites don't offer the full article on their RSS feed.

For example Ars Technica. I use miniflux, and it only shows me the first paragraph. Using the Download button in miniflux to download the whole article does nothing.

lire ios app can offline sync and can pull the original article instead of the rss stub. i use it with a local freshrss instance that i have accessible via tailscale

Yup. I’ve been using Miniflux for this for a while.

Or you can skip the route to online and just pull to your device at which point its cached for offline.

Never mind the fact that desktop apps for the birds. Microsoft won't even tell you what widget set to use post the failure of Metro, "use Electron" seems about as official as anything else.

From day one I planned to use my RSS reader on both desktop computers and a tablet (via Tailscale) and when I got a Meta Quest 3 I found it worked great on that although enlarging touch target to the AAA standard helped a lot.

On iOS, lire caches RSS-syndicated articles and images for offline reading. One-time purchase, no subscription. There's also OSS NetNewsWire.

lire is great. Ditched my Feedly lifetime sub for self-hosted FreshRSS and lire and I'm reading my RSS feeds again.

I built and use HeyHomepage.com daily for the same reasons. I love having a list of all kinds of websites (mostly 'dev blogs' from all you guys and gals!) which in it's essence is a list of bookmarks to homepages. I throw RSS techniques against it and it becomes my replacement for social media. I call RSS therefor 'Really Social Sites'. No ads and the quality of the content is completely up to me selecting the right websites to follow.

I see only one post from one feed at a time. If I want the next, I click a button labeled 'Next random post' and it gives me the latest post from yet another feed. I only get the first two lines of a new post and then read the full article on the website if I'm interested. There's something to say for paying hommage to a fellow internet user who put work in building a website... I want to read your post on your website or homepage.

Expanding on this list I also published (parts of) this list as a sort of blogroll, or shared list. This shared list is viewable and clickable in the browser for regular internet users, and downloadable as an OPML file so RSS users can import the websites that I like and 'endorse'.

Also expanding on this list, I built some functionality called Newspaper which automatically (instead of manually) checks some selected feeds that I deemed extra interesting. The different articles from different sources are than presented to me in a newspaper. Every time I log in there are some newspapers waiting for me. I'm always looking forward to the one called 'Cars'. The rest is mostly work stuff. ;)

It goes without saying that Hey Homepage is not only an RSS reader, but that it also has functions for your own timeline of posts with accompanying RSS feed.

The open web is not dead. You neglected it for too long. It misses you. Give it some love back.

I'm a GReader refugee that migrated to TT-RSS, and then FreshRSS. But somewhere in there, I wrote my own version of this same idea: pulling RSS feeds server side via cron every several hours and rendering a rollup. Never put the code online, but here's a paste of it if anyone is interested. I called it "Cooler" since it was designed to surface good water cooler chat topics. It's just a single Python script with a couple of dependencies.

https://textbin.net/kdhkz0nnyx

It tracks no state, but does color by source, and fades with age so you get a good feel of what site the post came from and how long ago. I had an idea to version control the static HTML after it was generated, so you could rewind time and see what the top stories were, but haven't gotten around to it.

If anyone cares, I'll host the full repo and share.

It looks like Dave Winer built something similar.

https://blogroll.social

http://scripting.com/2014/06/02/whatIsARiverOfNewsAggregator...

Unrelated, bit it's fascinating that the Google Reader API (which is super weird I think) is still the status quo for interacting with feed readers (https://freshrss.github.io/FreshRSS/en/developers/06_GoogleR...)

For some time I have been playing with RSS and web crawling myself.

RSS is not that simple and all common properties are read into a simple JSON data. Maybe somebody will find it useful.

https://github.com/rumca-js/crawler-buddy

> If I don’t remember whether I read the article or not, I might as well re-read!

Unless you remember after reading the first paragraph or two, in which case you've just wasted time partially re-reading them. This type of tracking is perfect for offloading to a reliable digital accountant!

> wasted time partially re-reading them

Refreshed low-latency cache for spaced repetition and creative working memory.

The thing is, the publisher are not obligated to show all recent post, some may show 1 some may show 5, but they might publish more than what they show in your refresh interval (e.g. 1 week) In this case, using a stateless reader, you will start to miss article.

Tell me you are JavaScript developer who likes to reinvent the wheel, without telling me you are a JavaScript developer who likes to reinvent the wheel:

- "RSS is for notifications". No, it's for content syndication. It is right there in the name.

- "XML is complicated, JSON Feed is better". Oh, dear Lord, forgive him for he has no idea what he is saying.

- "Lets ignore all the gazillion libraries for and tools for parsing and processing OPML, Atom and XML so that we can build a system that depends on deno a f*cking GitHub actions"

AFAIK author is not JavaScript developer.

I used to do frontend development ([1], [2]), now I am more of a database person ([3]), though it mostly is just prompting llvm to generate the right code!

[1] https://github.com/intellij-rust/intellij-rust

[2] https://github.com/rust-analyzer/rust-analyzer

[3] https://github.com/tigerbeetle/tigerbeetle/

> prompting llvm to generate the right code!

Hehe I almost missed 'v'.

Did you ask AI to make a roast? I'm genuinely asking.

It looks sensible to me. He's using two tools he was already using: Deno and GitHub. And he's using this RSS library: https://deno.land/x/rss@1.1.2

And he can always run that command without GitHub Actions if necessary or desired.

I've made https://rosselo.com exactly for being a RSS server side dashboard of news.

Feel free to try it out, it's completely free for now and upcoming future!

I remember using RSS heavily around 2007-2013 but is this still a thing?

I use RSS very frequently, it is often how I get my daily news.

Most blogs, at least in the tech space, have it. As well any major news publication worth their salt will have an RSS feed still.

Heck yeah! I use an RSS reader to consume news every day on both my phone and laptop. It's blissfully free of extras and distraction.

Using rn to access HN

Definitely! Advertisers hate it, because it's a way of basically bypassing ADs. I used to have my own RSS reader doing what the OP's doing, but I finally just started using Liferea (on Linux) which I love. There's a file format for sharing links called "opml", and here's mine for example (below). Many RSS readers can import/export your list of links to this format.

Each day there's about 150 new articles to scroll thru. What we need in the world however is some sort of OPML Sharing social media service where people can share their FAVs. It's a shame news sites are heading in the opposite direction with closed paywalls rather than openness, but I guess they're struggling to pay the bills. My apologies for posting such a big chunk of text and eating up half your screen. I only do this when I'm pretty sure it's relevant.

<?xml version="1.0" encoding="utf-8"?> <opml version="1.0"> <head> <title>Liferea Feed List Export</title> </head> <body> <outline title="Example Feeds" text="Example Feeds" description="Example Feeds" type="folder"> <outline title="News" text="News" description="News" type="folder"> <outline title="Ars Technica" text="Ars Technica" description="Ars Technica" type="rss" xmlUrl="https://feeds.arstechnica.com/arstechnica/index" htmlUrl="https://arstechnica.com"/> <outline title="Reddit - World News" text="Reddit - World News" description="Reddit - World News" type="atom" xmlUrl="https://www.reddit.com/r/worldnews/.rss" htmlUrl="https://www.reddit.com/r/worldnews/"/> <outline title="NPR World" text="NPR World" description="NPR World" type="rss" xmlUrl="https://feeds.npr.org/1004/rss.xml" htmlUrl="https://www.npr.org/templates/story/story.php?storyId=1004"/> <outline title="Simon Willison's Weblog" text="Simon Willison's Weblog" description="Simon Willison's Weblog" type="atom" xmlUrl="https://simonwillison.net/atom/everything/" htmlUrl="http://simonwillison.net/"/> </outline> </outline> </outline> </body> </opml>

I feel like you should post that on pastebin or something instead of posting such a huge piece of XML in a comment.

yeah, I shortened it just now. I mainly wanted people to see that file format, not the links in it.

The ads can be included in the feed, so no bypass here

Thus the "basically" qualifier in front of "bypassing".

I'm "basically bypassing" ADs simply because I never encounter them while reading my feeds.

thank you! I'd love to see all your valuable feeds really

My YOShINon RSS reader incorporates about 110 feeds, and, unlike the majority of RSS readers, actually treats RSS feed items as items that are stored in a database so it can present those items in various ways. Most news sources ranging from The Guardian and The Bulwark and Jacobin as well as phys.org and MDPI publishing still have RSS feeds.

It has screens to view things based on database queries, full text, and embedding similarities but the main UI looks like TikTok or Tinder for text, showing me articles it thinks I will like mixed with random articles to keep it calibrated. It spins like a top. I also have another thing called "Fraxinus" which was a cut-and-paste job from it that works as a bookmark manager and image sorter, the plan eventually is to mash them back together.

It sounds interesting. Is it available for download?

I was curious too so I Googled it. The hits were just similar posts in HN comments and this one on Tildes says more (assuming it is the same person):

https://tildes.net/~tech/1e5n/rss_users_how_do_you_use_organ...

A current priority has been to get it off ArangoDB onto Postgres in which case either an open source release or commercialization seems possible. And yeah, that's me. Shoulda been working on that code but I was playing video games instead last month

https://mastodon.social/@UP8/114740331853557527

Interesting idea, but it mentions the idea of notifications. Is the notification when happens to be new article is at the top on the page? Or is there an external service he forgot to metnion?

Reply on news.ycombinator.com