Unread items are marked as read after 30 days.

Can my auto-mark-as-read level please be turned off, or if that’s not possible, set it to 1000 years?

There’s a web comic I read, but after letting a few days go unread, I conciously decided to take a break from reading it. “But not to worry, NewsBlur will keep track of where I left off.”

It was when I noticed that the number of unread items remained constant at 30, I found out that this limit existed.

Some blogs I only dip into occasionly, leaving them unread maybe for months.

(I posted this as a comment to a related closed issue but seems to be ignored. As this is really a new issue I’m raising it as a new problem.)


It’s not a new issue. It’s by design;

  1. Sam is of the opinion that if you haven’t read it by the time it becomes automatically marked as read you didn’t care that much anyway (or you’re on vacation which is a separate real issue).
  2. It increases the load on the servers.

It took a great deal of convincing to get Sam to increase it to 30 days for premium members, up from the regular 14 days.

When changing it to 30 days he originally intended that time to be for all member but that slowed the service down too much (see avg. load time on the main page) since it adds more work for the servers.

It’s fairly safe to say that you won’t see this increased further and if you did it wouldn’t be by a lot.

You’ll have to adjust (like a lot of us did) or use something else, just the way it is sadly. I used to be pretty adamant that even 30 days wouldn’t be enough but here I am, still using a free account and not ever having items auto-marked as read.

Thank you hampus for distilling my thoughts on this. Most competitors have the same limitation. If you are technical enough, you can run your own hosted install of NewsBlur (see http://github.com/samuelclay/NewsBlur) and that number is easily tweakable.

1 Like

Okay, there has to be limit to how far back you’ll store the unread flags. I imagine it would take some serious reactoring to switch to a range-of-IDs based database schema.

How about, instead of sunsettig the unread flags after an abount of time, instead limiting to a number of posts? (Say 500.)

Consider two feeds, a daily webomic and a fifty-posts-a-day new site, both of which I stop reading for a while.

The news site will reach 500 posts ater 10 days, but the daily webcomic wouldn’t reach that threshold until I stopped reading for a little under two years. By that point I should probably unsubscribe.

I’ve not looked at the NewsBlur architecture, but I’d have thought the load of the server and DB would be proportional to the number of records rather than the amount of time they represent. So in this hypothetical world, the load taken by these two feeds on NewsBlur would be about the same. (Wouldn’t it?)

1 Like

There was no limit on Google Reader. None, zero, zilch. There may have been a limit to how many items it tracked (though I don’t remember one) but there wasn’t a limit to how old they could be.

Newsblur should not be just for new blogs. I found an old blog with 50 posts I want to read. The posts are all > 1 year old. Because newsblur marks them all as read I have no way to track which posts I’ve already read and which I haven’t. The posts are long so I can’t read all 50 posts at the same time.

Imagine a site that’s like each chapter for a book like Effective C++. There’s no reason only people that discovery the site the moment it opens should get to benefit from having things marked as read or unread. Even a user 5 years from now discovering a site like that would surely find it indispensible to know which entries they’ve read and which they haven’t

Google Reader did have limits, 1000 unread items tracked per feed if I remember. The difference is that they used actually copy the whole posts to their servers so you would never actually lose something. I had feeds in GR where the website no longer existed.

As long as they were full text feeds, that was awesome, but Google shut the service down because it was uneconomic even for Google. We will never know how much they would have had to charge to make it work because they decided against that option :frowning:

Newsblur is synced to the RSS feed, so how old the posts go depends on the author of the feed. The GR way used to annoy writers periodically because draft posts would sometimes end up hanging around for ever, not that I don’t miss it.

If I need to catch up on a feed like a web comic, I just set it to show all, oldest first. I manually mark as unread where I am up to. If the author have removed the old posts, then that’s life I guess.

I think you missed my point entirely.

I can not mark old posts as unread period. It tells me “post is older than 30 days, can’t mark as unread”.

So, I find some year old blog with 20 tutorials on WebGL. I want to read them. I want something to track which ones I’ve read. Google Reader would do that. NewsBlur will not.

That’s a legit use case IMO. It’s sad that NewsBlur can’t handle it.

As for Google Reader, Google did not decide it was uneconomic (I worked at Google). Google decided they wanted those programmers working on other stuff. Google has cancelled plenty of things that made them money. They cancel stuff because they have a limited amount of resources (people) and bandwidth (time to manage) and want to use them for something else.

You are right on the mark as read, sorry.

You can “Save” a story that is older than 30 days (and unsave it), so you can mark the next one to read temporarily.

Maybe a number of unread posts limit per feed could be an option rather than number of days.

Johnny, Google removed engineers from Reader at least two years before it shut down, bringing in a handful to deprecate social features in October 2011. They didn’t sunset Reader because of resources but rather “focus”. Also, there was a big database change and Reader needed to go through a lot of work to continue, so that was when the decision to shut it down was reached.

My source for this is having spoken to a handful of Reader engineers and others over the past few years. Who knows what the actual truth is. I agree it was a bone-headed, short-sighted move, not that I’m complaining.

Here is a situation that makes this limitation really dumb. I wanted to use newsblur as a podcast reader since I refuse to install shitty Apple software, there’s about 420 episodes I want to listen to, but since it won’t let me mark as unread all the way back its useless for this. Should be an option for paying users.

By the sounds of things this will never change. I don’t know the details but it sounds like it would be a massive resource cost server side.

If you don’t want to use iTunes I’d suggest Pocket Casts, they have a web version (one time cost). Can’t think of any native RSS or podcast app to recommend, very few options for Windows (which is what I guess you’re using).