Library clips

Feedeye is doing what I see as the next step in RSS Readers, this is clustering stories, plus ranking items by popularity.

I’ve explained in my Google Reader post how clustering stories and ordered by popularity akin to Google News or Techmeme saves time and just feels right.

The usual scenario is you finished reading your RSS Reader and reflected back and noticed that 20 of your feeds posted about the latest daily breaking news eg. Google aquired YouTube. These 20 stories were scattered in your RSS Reader, so you read these similar stories at different times.

Naturally your RSS Reader would be smart enough to group or cluster these stories for you so you can read them one after the other.
The first on this scene at a personal level was Feeds 2.0, followed by WizAg, and MyFeedz.

Both megite and TailRank will also do this, but these aren’t personal RSS Readers where you can mark a story read/unread, plus these two services don’t show every story in your feed set, they just show the popular stories with the clustered items, plus they nudge in some recommended stories.
I like the idea of Feeds2.0 and Feedeye as they will cluster stories, but you also get to see all items in your RSS reader if you want to keep reading, and you can keep stories unread.
NOTE: Feedeye has different views, some views don’t allows to keep stories unread.

The main difference between Feedeye and Feeds 2.0 is that the stories in Feeds 2.0 can be ranked by date or past reading behaviour, whereas in Feedeye the stories are ranked by popularity.

Personal meets popularity

What I’m after is a combination of popularity ranking of stories and personalisation ranking of stories (plus clustering similar items).

Feedeye displays stories in order of popularity and the stories with clustered items rank on top.
Feeds 2.0 ranks stories by personalisation (based on past reading behaviour), and shows clustered items to stories where available.

What I’m saying here is that I still want clustered items, but I want stories ranked according to me, not just how popular everyone thinks they are.

In fact, I want both, show me two streams:
Stream 1 - stories ranked based on my attention
Stream 2 - stories ranked by blogosphere popularity

NOTE: Any clicks in stream 2 would not track reading behaviour.

Maybe you could make one stream more prominent and the other stream in a small scrollable box.

And what about stream 3 - a mix of both streams.

What this enables:

1. I have muliple views of a story all bunched together (clustered), I can be informed from different sources about the same story in less than 5 minutes (according to the feed sources I have chosen).

2. In one stream stories are ranked by my personal click behaviour
- the idea is, as I read my river of news and start to come to stories that don’t interest me I could feed safe to not read the rest of the news items, and just mark all as read, this is an ideal world
- option is to read the remaining items by tags as in MyFeedz.

3. In another stream I can read stories ranked by popularity
- a story is popular if lots of people are writing about it
- a particular story is popular (the mother of the clustered items beneath it) if all the other stories link to it, or it is just well written, or if it broke the news first, or if it is considered generally popular from past stories this source has published, or a mix of all this.

Also to mention with all these services is that they throw recommended items into the mix, so for every story in Feedeye, you can show related items from the blogosphere. This is a great idea as you don’t want your feed set to work against you, it’s refereshing to see what others (feeds not in your feed set) are saying about a story.

What is popular to you?

Feedeye is a personal Google News or Techmeme, you get to see every item, but the popular one’s are on top.
In fact all stories are ranked by popularity, it’s just the stories with clustered items are on top.

Why are they popular?

Because there are several posts in your mix talking about the same thing, so naturally it is popular.

BUT…this is still not enough for me, as I might not care if they are popular.

I need my river of news ordered (stories need to be ranked) by learning from my reading behaviour, but I still want clustered items. It’s just these posts with clustered items don’t necessarily have to be at the top of the pile.

This way I’m reading what is personally important to me first, and when available please cluster similar stories together.

Feedeye

Add a bunch of feeds or an OPML file or OPML URL and give it a name, this is called a set and gets its own public URL.
A set basically would be the same as a folder in your RSS Reader, a bunch of feeds according to a topic, or whatever.

You can also filter text in the set content: filter out in/out/highlight terms in the body, text, author, URL…

NOTE: Not sure if it has deduping (if some of your feeds are news search feeds from different news engines you would get identical items appearing…I don’t know anyone who would want to see the same item more than once).

Sort stories by date or grouped by number of similar stories
- even though you sort stories by date rather than popularity, I don’t see why you still can’t have clustered items, not sure if this is the case.

Various reading view - 2 pane, newspaper, flowed summaries, slideshow (quite unique)

Separate view for stories you have read (clicked)

Temporary separate view for recent items (only on 2 pane view)

Keep a story unread (only on 2 pane view)

Clip stories (this is saving or bookmarking),

See related items from the blogosphere.

Add a button on your blog for people to subscribe to your blog in their Feedeye account.

Each of your sets has a public URL, so basically you are making your own Techmeme or Google News, that is, memetracking according to your feed sources.

NOTE: maximum of 40 feeds per set, and a maximum of 25 sets.

Memetracking or related items or clustered items, whatever you call it, is so great for journalists and consumers as for any feed set, you can see the different views/published stories about the same story.
You can read a story and before you wonder what anyone else is saying about the story you have already clicked on it, and on another, and other.
In 5 minutes you have been able to read 5 news items about the exact same story.

Plus if you only have 15 minutes to read stories, you can read the popular stories according to the world.

So in a short time you can get your daily digest and not stress to much about missing anything, this could become the opposite of RSS overload.

See my Feedeye for my Top 20 feeds (this is set on the 2 pane view):
http://feedeye.com/johnt/top20

Here is re-posted some of the Feedeye FAQ, it just explains this concept so well:

“Feedeye is an online feed reading service (sometimes called an “aggregator” or “RSS reader”), much like Bloglines or NewsGator. But Feedeye can do more than just show you your feeds: it’ll also tell you which items on the page are the most important, introduce you to cool new feeds you mightn’t have otherwise found, and let you easily browse what other folks are reading.

Feedeye also has a bunch of other features, such as the ability to change what your sets look like (to a slideshow, or a newspaper-style page, or a big long page of new items, or a traditional feed reading view with two panels). Unless you mark them as private, anyone can visit your sets or link to them from other Web sites. And Feedeye can help you discover new sets similar to the ones you’re reading.”

And more:

“The main objective of Feedeye is to give you some sense of what’s important among the items in your feeds, rather than just showing all of the items equally as normal feed readers do.

This means that Feedeye provides the traditional function of a newspaper — giving you cues as to what’s important — but lets you choose the sources that will determine importance. It’s a bit like creating your own newspaper and appointing your favourite blog authors as its editors.”