home / blog / notes / photos / about / more

(RSS) feed metadata syncing

From Google Reader to self hosting

Posted by Jeena

I have been using RSS and Atom feeds for nearly a decade now. It got a bit less while I was mostly using Twitter to get most of the news, but since Twitter is fucking with the twitter client developers and users I almost stopped using Twitter and went back to the good old RSS/Atom feeds.

And it is great, I use Reeder on my Mac, MobileRSS on my iPhone and Google Reader on my Android tablet.

But there is one thing I don't like about how it works right now. Every one of this clients relies on the undocumented Google Reader API to synchronize the "read" marks and the subscription lists. The problem with that is that Google doesn't really want this users any more and they would love all of them just to use the closed and centralized Google+ instead for this kind of stuff.

I don't know that but I suspect that Google will at some not that distant point in future start discontinuing Google Reader and the undocumented API. Take a look at their Google Reader Blog for example, in the years from 2005 up untill last year it got updated almost two times per month with new interesting stuff happening and so on. But since october 2011 not a single update, and that is right after banning a buch of stuff people have been using and moving that stuff to Google+. That is obviously ok for them to do, it is their application they can do whatever they think is best for them to do.

But for users like me who only use Google Reader for synchronization between their desktop and mobile clients across different computers, iPhones, tablets and so on, it is a bad thing, because it is the beginning of the end. The only thing even worse than this is that even though RSS and Atom are open formats, there is no open protocol which this clients could use to synchronize the "read"-marker or the subscriptions.

What I would like to see is some protocol where I then could implement and host my own server which would save the read/favourites/etc. markers and my subscription with folders which clients like Reader, MobileRSS and NetNewsWire could use to synchronize this information across them.

It could kind of work like the undocumented Google Reader API but honestly it has so many flaws like for example only one level of Folders and so on that I'd be ok to just have some very light protocol for synchronizing metadata about "read"-marks and subscriptions, everything else like getting subscriptions from the servers can be done by the clients themself.

I have been thinking about it a couple of years ago, and even started designing some REST API for it, but back then it seemed not to matter to anyone, perhaps today there will be more people interested in something like that?

[update] After I published this, I got a couple of things for further reading:

1 Mention

Have you written a response? Let me know the URL:

There's also indie comments (webmentions) support.