Sit down and really write out your requirements.

Two things will emerge from that exercise: 1. there probably is in fact a reader out there that meets it, you just haven't found it yet or 2. your requirements require solving the AI-complete problem of perfectly determining what you are interested in.

Generally I find most people slot into #2. I am not kidding. I've been listening to this complaint for about 8 years now and every time I've worked with someone, it's one of the two above, and usually #2.

A third possibility is that you are just trying to read too much, period, if even an AI-complete filter would still leave you with too much. The suck may not lie with the software, it may in the requirements.

The problem of reasonably determining what someone is interested in from what they've previously read, skimmed and skipped shouldn't be "AI complete."

It only involves matching patterns to key-words. It might not be easy and the results might not be perfect but it seems possible. Using data from friends and finding the best correlates could help too. You could use the Netflix algorithm or other magic.

I suspect that's why today's readers don't satisfy - they aren't there yet and it's easy to imagine they could be.

I agree with your claim that recommendation algorithms aren't AI complete, but for different reasons.

"It only involves matching patterns to key-words." False. Recommendation might be far trickier.

Nonetheless, if you can do recommendation perfectly, it still doesn't mean that you can solve NLP, machine vision, control (robotics), planning, or any of the other major AI tasks.

Well, I was speaking broadly...

But the point is that we have many examples of recommendation systems that do work reasonably well and working reasonably well is what I suspect would satisfy most users.

Further, a reader which could learn reasonably well from passive observation would give users the base do explicit tagging and rating.

Implement it. Prove me wrong. Sell it and get rich.

Good luck.

I think I created a very nice feed reading experience with NewsBlur: http://www.newsblur.com.

It shows the original site, allows you to read as you normally would, but keeps track of the stories you're scrolling past.

It also allows you to filter stories based on what you like and dislike about them: words/phrases in the title, tags and categories, authors, and the publisher themselves. There is a slider that allows you to show/hide stories based on this filter. It's very fast, too.

I am writing an iPhone app so you can use NewsBlur everywhere. It's just a hobby project, and people have so far been impressed. But I would love for NewsBlur to become a useful tool that people choose to use.

I wrote it because I was also dissatisfied with readers, especially Google Reader. I also knew Python (Django!), JavaScript, and wanted to put them together to test my abilities.

Most importantly, I'm writing it all as open source. The iPhone app I'm writing, the feed parser and fetcher, and the front-end reader are all on GitHub: http://github.com/samuelclay/NewsBlur/

Awesome work! I'm definitely using this over Google Reader from now on!

But there are few tiny details I hope will be fixed soon: The "Original" tab is the default one - which slows down the whole thing if I open some content-heavy site, and the reason why I use feed readers in first place is not to have to watch those horrible sites at all, so it would be nice if we could change the default view.

The other thing is that there is no way to change the folder to which certain feed belongs. As well as someone before me commented, you can't delete feeds either.

The default view is the last used view for that feed. I'm planning to have that in later today.

Drag and drop moving feeds is coming sometime this Summer, but you can delete a feed right now. There is a "manage" button at the bottom of the feed bar.

Awesome reader, first one I've actually liked.

Only major problem so far is that I can't delete a feed (the link isn't working) or edit details.

Other than that, awesome job.

Thanks a lot! I'll fix the link pronto. :)

I haven't tested it, but I'm intrigued by Shuan Inman's Fever reader (http://feedafever.com/). Basically the idea is that you separate your feeds into two distinct types: essential feeds that you read completely, and supplemental, high-volume feeds (aggregator-type blogs, like Engadget). Then the essentials are cross-referenced with the supplementals to see what you're likely to be interested in.

I have a similar strategy, but instead of trying to manage the "supplementals" in my RSS reader, I try to a) identify them, and b) banish them back to the browser where they belong. Identifying them can be tricky, at least for me, because I tend to think that feeds I've been reading for a long time are essential, when often they're not. Removing a lot of politics blogs from my reader I swear has made me less stressed, and I don't feel any less informed.

I think the higher volume of content means we have to become better at selection — not only in choosing what we read, but also in identifying and not reading what is noisy and lacking in substance, even when tempting (ahem, HuffPo). Edit: and I'm not sure algorithms are the solution; I think it's something our generation will have to learn — to be our own conscious curators.

I'm using essentially the same strategy. Basically I split feeds in two categories:

"Fresh" ones I either read on the spot, or mark-all-as-read. These are the ones I usually check frequently, and very quickly. And I don't have to feel bad about marking all as read :-)

"Evergreens" are ones I like to read entirely, but not necessarily now.

I'm using Google Reader, but these days mostly through the splendid Reeder on the iPhone (http://reederapp.com/).

I really like Reeder for the iPhone. Google Reader fluid app on the desktop and Reeder for the iPhone works great for me.

Reeder for iPad is great, too; really fast UX.

I'd use Reeder if it weren't for the fact that I read a bunch of math blogs and Reeder usually doesn't want to download the PNGs of formulas produced by the LaTeX plugin for Wordpress. It's OK on other feeds I've tried it on.

Great UI, et al.

I also like that it aims to help the user make sense and filter their news sources. But you completely lost me when you were explaining in the video how you edited the settings to see Comics. I'm confident I myself could figure it out with time and effort, but it goes over the heads of the common man, which I believe is who we're talking about now.

For me, it's a BIG drawback that it's not a hosted service. I don't want to be updating my feed reader myself, I don't want to have to install it and deal with server config issues that might come up, I want to access it from anywhere (I'd probably host it locally on my network, otherwise) I want the computer hosting the service to do all the grunt-work.

I use Fever for my feeds, I simply love it. I tried others but this is definitely the best one I found so far.

I use and love Google Reader. I use the List mode (as opposed to Expanded) and my work flow is the same whether I'm at my desktop, iPhone, iPad, or Blackberry.

A handful of feeds I insist on reading everything, which means I hit J-J-J-J to get through them all. But most I merely skim the unread titles for things that look interesting, and when I'm done I hit Mark As Read to clean up the rest.

I like how I can do this at feed-level, category-level, or with all unread items at once.

And if I'm mobile and I want to see something from my desktop, I star it.

Admittedly, the last time I considered using a thick-client reader was way, way back in 2004 when I was still Windows-only, and I wasn't happy with anything I saw. I've used Google Reader since it was released in 2005, and I've had no reason to use anything else.

The other advantage to Google Reader is that the processing is offloaded. I have many active feeds, and when performing an update my local client on my 8GB quad-core raid-0 system would grind away - no matter what client I used. With Google Reader, it's Google that is doing the data collection and processing, making it light-weight enough for anything that can run a browser.

Same here. I have a folder for must-read feeds (mostly close friends and a few VERY high-quality, low-volume feeds) then I categorize the rest by interest. I've also leveraged my friends who filter out the relatively few rare gems from insanely high-volume feeds such as BoingBoing, Engadget, RWW, TechCrunch, etc. It's awesome. I don't subscribe to them, but I still get the highlight reel on the good stuff because I network with like-minded folks on GReader.

I do monitor some high-volume feeds (HN among them) and I skim the feed then share the good stuff that I think my contacts will like. I star things that I want to revisit and permanently bookmark stuff that's highly relevant to my work or hobbies, genuinely awesome or fascinating.

I wondered this myself until I hit upon the simple solution of deleting feeds with too low a signal to noise ratio (for my personal tastes). Expecting RSS readers to tease out relevance means having to give them an idea of what's relevant to you. This seems to be a self-inflicted problem on the part of people who use feed readers. I've tended to think I "needed" to follow more than I could possibly consume and retain. Pruning my feed list not only helped with overall comprehension of feeds I'm interested in, but eased a somewhat bothersome feeling that I was falling behind on a keeping up to date.

Personally I ditched NetNewsWire after Reeder for the iPad came out. I have Reeder on my both my iPhone/iPad and just use Google Reader for those times I check feeds on my notebook. The user experience (note that I didn't say user interface) it provides is astounding. Navigating feeds, marking read/favorite, and the article typesetting/layouts are all the best I've seen in an RSS client to date. Combine something like Reeder with a carefully curated feed list and I don't really see there being much of a problem.

I've had a similar transition here - I've pruned almost all high-volume blogs from my Google Reader, and focused more on people whose opinion I value very highly, Twitter, or social news sites (HN/Reddit/etc.)

So far it's worked very well. I still seem to find out about everything I'd care about, but have a post volume on Google Reader of around 60 posts a day. I star anything longer I'm interested in for later reading on iPad/iPhone (either through Reeder or Instapaper).

I probably sort through several times more tweets, but that's something I'm far more likely to do when I have 30-60 seconds - so it's spread out throughout the day, and anything longer winds up in Instapaper for me to do at a later date.

In short: Try social filtering. Find people who share interests and blog/Tweet/etc. about them, and let them be your filter.

I have a similar setup. Add anything I am interested in to Google Reader, consume via Reeder on iPhone/iPad, or consume via Feedly if using a browser.

Reeder and Feedly both rock, but the key is eliminating low signal to noise sources.

The most obvious thing is that they don't do enough to help you filter the wheat from the chaff. The sheer volume of news that's available is too much for anybody to consume... it is almost literally like trying to drink from a firehose.

So our readers need to do a better job of helping us figure out which stories are relevant/interesting to us, give us ways to skim and summarize without reading the whole article, filter out "dupes", etc.

The problem is that unless that filtering works magically in the background, it just adds interface debris. And it's easy to reach a threshold there, where just scanning for the stuff to filter (hidden amongst the UI elements) and you might as well be better off doing your filtering yourself.

Which right now works best for me. If I use Google Reader directly, it's set up to use the re-reader style, at home I use NetNewsWire with the subscription list hidden. So to browse through it, I can mostly use keyboard shortcuts, so skipping something is easy. Reading is comfortable enough, but if the article is too long, I either open it in a background browser window or send it to Instapaper.

Works well enough for me, although if I had more newspaper-style story listings in there, that just contain the headlines and link to the main story, filtering and adding to the queue might be more difficult.

And alternative would be some kind of social filtering, but that's close enough to HN or Reddit or just listing the links you've got on your twitter inline (c.f. http://news.peepcode.com/)

I use Postrank (formally AideRSS) for this reason - BoingBoing have far too many posts for me to read, so I run their feed through Postrank, and get only the most interesting posts.

http://www.postrank.com/main - their index page is baffling.

A friend recently launched http://feedingo.com/ which basically tries to make the user experience as friendly as possible.

Pricing is the only problem, but after getting feedback he plans on trying to address that promptly (like $5/yr or something, for a non-free version)

I'm that friend! I had the same feelings about RSS as the OP, so I created http://feedingo.com.

Most RSS readers try to be something they're not: "it's like reading email", "it's like reading a magazine" etc. Feedingo is a web app with an interface that makes reading RSS as easy as possible.

It's true that the current pricing that I launched with needs to be changed, but sign up for the free version and try it out while im working on that.

I think you just need to think about what you want, and think about what sources you read from. You say you have 68 sources in your feed, whereas I have ~100.

Personally, I like Google Reader, but I don't use that to read my RSS (just manage it). Instead, I use Google/ig. I have several different categories setup and I don't necessarily read everything. One of the categories I have setup is mRead (short for mark as read) and I don't really read those articles (often), but I still subscribe incase I see a headline that screams my name, or if I want to use Google Reader to do a search within my subscriptions for something specific.

If you have too much noise in your RSS versus signal, than you may want to double check what you've subscribed to and cut back, or add a "mRead" category of your own.

I think there's an issue with people's expectations of their Feed Readers. If you add 100+ feeds then expect the reader to make it easy to read every single post, that will never happen.

Feed Readers work best, and are much more realistic when you treat them like Twitter. You can follow 100+ people, but you don't expect to read every single tweet that they send out. If you miss a post, oh well, it's yesterday's news.

Just log into your reader of choice when you have free time, read the latest news, then go about your day guilt free.

plug: http://feedingo.com

Free tip if you want to write a better RSS reader: research Usenet newsreaders like trn or tin. Many of the problems with RSS that everyone is grappling with now were solved (to varying degrees of success) decades ago. Usenet news and RSS aren't 100% analogous of course, but they are similar enough that there's a lot of man-years of thought which can be reused.

I think it's not the RSS readers' fault. It's the whole reading on internet that sucks.

P.S. - http://goodnoows.com did look promising. Found on HN recently.

I'm not sure if the question is aimed at geeks (ie HN crowd) or in general, but my own observations from previously working at a leading news website indicates that people find the lack of visual aesthetic disappointing in News Readers.

Images are often not included in RSS feeds, and if they are are then they are only shown once you are viewing an item.

When we are triaging which stories to read on a news/blog site, imagery plays an important as we scroll and select - and I don't think I've seen any RSS readers which show you a feed item's images at the story selection level.

The lack of site aesthetic/look and feel maybe less important to geeks but end-consumers also miss that when they are viewing news stories in a reader. In other-words, when I read CNN's feed in my reader I don't get any of CNN's site aesthetic.

Finally, don't forget comments and also the serendipity of suggested related content -- both of which are also missed out when we read stories in a news reader.

Check out Yahoo Pipes. This under-appreciated tool can help you be more selective in the news you actually look at.

http://pipes.yahoo.com/pipes/

I used it to make an RSS feed which only shows stories that make it into the top 10 on HN. http://pipes.yahoo.com/pipes/pipe.info?_id=91d82e993e4607b9f...

If you're willing to forgo graphics, I find newsbeuter (http://www.newsbeuter.org) to be a great reader.

It has excellent mutt-esque key bindings, amazing customizability, etc.

I use Feed on Feeds, a PHP script that runs on my home machine so I can access it from anywhere. It pulls in feeds once an hour and I view them in a big list. It's got a bookmarklet that makes it easy to add a new feed when I'm looking at a site I like.

My workflow is to scroll down the list of unread feeds. If something looks interesting, I either read it within the browser window, or I open the article in a new tab and continue scanning. When I hit the end of the list, I "mark all as read". If I don't finish the list I can mark everything up to the one that's visible as read. Then I go through the tabs I've already opened.

I follow about 200 feeds, most of which are infrequent posters, so I'm going through about 100-200 items a day. Out of those I find I open up maybe 5 or 10. Since I'm scanning everything in a row I can go through the list in a few minutes if I check a couple of times a day.

In most cases I've found that high volume feeds have the good stuff duplicated elsewhere, so I can unsubscribe from the high volume one (though for some reason I still subscribe to Slashdot, even though I learned about all the good stuff a few days prior from other feeds)

The Feed On Feeds interface is also so sparse it doesn't get in the way.

I wrote myself an RSS reader that does exactly what I want...which is aggregate links for me.

It's here: http://newsyndicated.com if anybody is interested.

The only real "feature" of it is the ability to make a sort of "meta rss" feed so that you can share links that you have "liked" with your friends.

I think you need to categorize your feeds by importance. There are some sites you will surely want to catch everything from. There are others where you do not. Do not categorize your feeds by topic but by importance.. that way you can read the "Must Read" folder every day and hit mark all as read on the "Less Important" (or similar) folder if you haven't got the time. That's how I manage my Google Reader and it works a treat!

That aside, relating to filtering the news, my site at http://coder.io/ is doing exactly that but with only coder/programming related news. Still in alpha and very early days but I'm hoping it will resolve the problems you raise for one tiny niche.

I would say it's a combination of a number of reasons, some -- if not all -- relate to the technologies encapsulated in RSS.

First, because RSS by design is meant to just be metadata, and because it's really hard to render HTML (especially those that are not well-formed) the readers typically have to jump through so many hoops to get things just working.

Second, I think it's also because the RSS readers are typically developed by non-designers. Although Google Reader keeps it simple, it's utilitarian and mostly not very pleasant to look at. I've tried others (which are also free) but typically it doesn't come close to making it enticing to read.

Maybe someday someone will do a feed reader that renders articles in a sane and readable way.

Hey dean, give my reader http://feedingo.com a look. As an example of an RSS reader created by a designer.

Does anybody know if there's an RSS reader that allows regular expression tagging across feed content? Eg, show me all RSS posts that contain "iPhone" or that match the regex "(?:iPhone|iPad|iPod|Apple|Android)" etc, as a simple example.

Try Fever: http://feedafever.com/

I love Fever. Putting the high volume news sites as "sparks", and the high signal feeds that I like to read everything from as "kindling", gives a great way of seeing trending topics when you have the time, but without seeing a giant unread count in your list.

Well worth the one-time fee and setting it up on my Linode box.

It is a great idea, though a little too over-designed. The concept gets in the way of reading more than it should.

Is it the RSS readers that suck or the content itself? If you can avoid these sites that just aggregate content or provide second and third hand accounts of the same items it's easy to trim down your reading list.

I know it sounds like an ad - because it really is - but http://earlyedd.com wants to be the solution for the RSS problem (http://blog.earlyedd.com/post/739958248/the-problem).

EarlyEdd is the RSS reader that shapes on your interests. It discovers people reading pretty much the same stuff as you do (the fellows), and it compiles a relevant news edition especially for you. Give it a try ;)

Most RSS readers suck because they a) are too slow and b) assume you want to read every article i every feed.

I've not tried all the suggestions here (Reeder and Fever sound interesting) but having used GReader and Feedly I keep coming back to using Firefox Live Bookmarks with the LiveClick plugin. I can very quickly see and scan what's new. And with XMarks the feeds can be synced between your computers. Horses for courses I guess.

I've tried several Firefox plugins, Outlook, and a few others that I can't remember and finally settled with Google Reader.

Not sure why you think it's horrible. It's always available from wherever I am, mobile or desktop. Adding feeds is as easy or easier than most other readers I tried, and it even creates feeds for sites that don't have them. Also, the sharing and commenting features are nice.

Agreed. I put up a side project at http://beta.crunch3.com, which was an RSS-on-twitter concept. The interface is "slidey", similar to Pulse, but it gets relevant feeds from twitter, and takes screenshots like Google Fast Flip. I didn't continue development on it, so it's a little slow & the server will probably melt.

I recently launch http://www.gmbhnews.com as a mobile rss feed reader, give it a try and let me know if it solves your problem, all you need is a webkit broser for a better rendering.

For the desktop, you could try http://www.mcsquare.me , it should be fine as well.

Some readers cannot detect duplicated posts.

Well, especially when the same news story, say Yahoo getting $37/share from Microsoft, is reported on by TechCrunch, Mashable, TNW, RRW, SAI, etc.

This is very tricky to do. Let's say that it's the iPhone4 launch day and every story is about that phone. Some users might want all those duplicate posts, to get different perspectives on the story. Others might just be annoyed and want all those duplicates skipped.

I think filtering is probably the way to go in this case. Instead of detecting duplicates the reader should allow you to filter out all stories tagged "iPhone4".

I'm reading this in NetNewsWire for the iPad, a beautiful, slick, polished app that syncs with Google Reader. I have no complaints.

It sounds like you're subscribing to too many feeds. 2500+ items? You need to do an audit, cut that figure in half, then see if you're still unhappy with your RSS app.

After trying a lot of choices TwitterTim.es is my favorite. It creates a personal newspaper from links popular among friends-of-friends of my Twitter account. It produces very very relevant results for me.

I have a quite good experience with feedly. It's a firefox extension which integrates with google reader. The integration seems to have some bugs with new "shared items", other than that it's perfect.

Feedly rocks, I prefer the magazine like appearance to reading feeds (especially when I've got so many).

Wow - I love mine (FeedDemon: http://www.feeddemon.com/). So great - does almost everything I've ever wanted

Different people use RSS in very different ways.

Reeder on the iPad is spectacularly good.

Perhaps you could give some reasons why?

Immaculate UI/UX. Whether in landscape or portrait orientation, there are many clever flourishes: it uses Tweetie's "pull up to refresh" mechanism for moving between articles in a feed; or you can swipe an article left or right to return to the feed index; or you can pinch the article/feed to return to all of your feeds. In a feed index view, you can swipe a list item to the left or right to either toggle the read state or the starred state.

The feed index is designed very similarly to the iPad Photos app: you can either pinch open a folder to view the feeds within it separately, or you can tap it to view all of the articles in a combined view; similarly, you can pinch open a individual feed to preview of the articles within it.

Additionally, Reeder caches articles and images for offline reading; and the top-level feed organization is pretty sensible, too... it's separated into 3 sections: read, unread, and starred. It's also extremely fast, having been rewritten to use SQLite instead of CoreData.

What you said, plus it looks really, really nice.

IMHO it's the lack of obvious manual control over the special sorting algorithm that causes readers to suck.

Do we need an open source techmeme?

Because they are (mostly) web apps.

Www.Netvibes.com is my favorite.

RSS readers are fine, Google Reader is better. The problem is in user experience and how they use it. If you do not prioritize and constantly worry about missed or unread items, you will only end up blaming your reader. The trick is in organizing, prioritizing and ignoring the unread/missed items. With these three strategies, you can effectively absorb unbelievable amount of data on daily basis.

I have A, B, and C tags ordered by "importance" to me - i.e. mostly rss feeds where I want to read or skim every entry.

The rest are ordered by category/tag.

Works fine for me!

Also I use Postrank to turn some of the high-volume into low-volume high-importance feeds.