Hacker News Reader

Comment by themgt | Hacker News Reader

I think a 2013 Reader should be an open source web app you can run yourself or pay to use as a service, which runs on a server pulling in feed data, and then as a detachable HTML5+JS+localstorage client that downloads the data and can run offline.

Building another hosted RSS reader project just gets us back to step 1 in terms of one company's ability to yank away a tool that functions almost like a newspaper to many people today.

NewsBlur looks like it just got a ton of new github watchers since this announcement, and would maybe be good to treat as the defacto "open RSS reader platform" leader - https://github.com/samuelclay/NewsBlur

View on HN

As someone who works on an RSS crawler day to day, this is much more difficult than you think because of how poorly most people implement RSS.

If everyone stuck to the standards and did things the same way, it would be easy. If you can't account for these errors your reader wont work with 80% of feeds, and nobody will use it.

Rss feeds are usually programmed more poorly than a website itself, because it's one of those features that comes from a discussion that usually goes: "well everyone has an rss feed, so we should have one, so make one". Yet they know people will rarely use it.

You could make it work for the popular CMS's but even between those there are standards problems. The majority of the problems are in storing the stuff that's crawled.

> "how poorly most people implement RSS."

^upvoted! I agree, it's horrible! Not only are the many different versions of RSS & ATOM a problem, but the freedom of putting any kind of data in any of the fields. I've seen "descriptions" in the "title" property, "images" in "links", "links" in "images", "links" as "urls", "UUID" as "ids", and so on and on and on. If it wasn't for Java ROME, I would be spending dozens of hours just making our aggregator compatible to any newsfeed our users come up with.

Stick to the standards, folks!

We should really be ditching RSS in favor of more flexible platforms like those on schema.org. The web is constantly evolving and changing; to try to keep some extremely legacy tech in tow is just asking for trouble. It's not even useful for its structure, as people are misusing it, so I don't see why we should. At least following schema.org conventions aides in your SEO.

> an open source web app Maybe as part of an OwnCloud: https://owncloud.org/

("OwnCloud" is hipster for "web hosting", I guess?)

OwnCloud is a WebApp that replicates a lot of cloud services (file sync [android and desktop client], contacts, calendar, streaming music, etc.). It's available as FOSS or with commercial licenses.

And a replacement for Google Reader, ownCloud News, is under development.

http://algorithmsforthekitchen.com/blog/?p=479

Looking forward to it maturing. I have the alpha installed on my server but it's obviously still pretty buggy. They way it looks, all it'll take to bring it to Google Reader awesomeness is some CSS editing.

And it will sync with Akregator, sweet!

Super nice!

I've seen a lot of people bemoaning the loss of Google Reader's ability to see old items which aren't in the RSS feed anymore. Unfortunately, that feature can't be implemented as well without centralization.

I think it would be possible with a freenet/bittorrent like architecture, given enough users.

So you're saying that we should either be willing to pay $5/m for an RSS reader service, or $5/m to host the service ourselves?

Someone with a lot of feeds wouldn't be able to fit all of them in the 5 MB that DOM Storage allows. I wouldn't like an offline feed since a big majority of my feeds are videos and podcasts.

As for them being able to yank away at any time, you're right, but there's going to be enough new stuff out there that it wont be a problem, this time we wont have a monopoly.

> So you're saying that we should either be willing to pay $5/m for an RSS reader service, or $5/m to host the service ourselves?

Why is that so unreasonable? Though, I think it's likely that the hosted solution would be much cheaper than self-hosting, since you don't need to rent one server per user, a hosted solution could be much more efficient.

The ability to detach the frontend and control the storage backend was a pretty big motivation for http://grimwire.com/local/. You might implement the RSS backend to serve an environment for running RSS apps. Then you'd write the GUI as one of those apps. The separation would allow you to change backend or frontend independently.

I like this. Kind of reminds me of the WordPress.org/.com model. If I don't want to run it myself I can easily use .com, but if I want full control I can easily run the .org software on my own server.

Looks like this is a feed reader that I can run myself: http://feedafever.com/

Not free though.

I don't get it. You have a list of rss feeds. stick them in another client. What am I missing?

Reply on news.ycombinator.com