Google Reader needs clustering and more
Steve Rubel has pointed to a post by Curtiss Thompson on his Google Reader wishlist…the issue Steve is highlighting is collapsing similar stories.
We see sites like tech.meme and Google News collapsing similar stories, so why shouldn’t we be able to do this within our RSS Reader, after all our personal RSS Reader is like our own mini Google News or tech.meme, by that I mean we choose the news sources, ie. your subscription list.
This is exactly what Feeds2.0 allows you to do, and Google Reader should take a close look at this, as many power users are turning to Google Reader.
What Feeds2.0 has is kind of like memetracking, perhaps better described as clustering or collasping similar stories together.
How clustering helps a blogger
Just say you have 1000 unread items in your RSS Reader, you boil a cup of chai and get into rss zen mode.
You come across a really cool story you want to blog about, but before you publish a post you want to see who else may have posted on this same story, so you cruise around your RSS Reader clicking on your favourite feeds to see if they have also reported on the story.
Once you have collected a few links you decide that’s enough, you are not going to wait and read every item in your RSS reader before you post this story…if your like me I never get round to reading every item.
So you decide to publish your blog post.
Then you continue slogging through your RSS Reader and 1 hour later you come across a similar story that you could have included in the blog post you just published, and half an hour later you come across another similar story.
Damn, I can go and update my published post, but you would of prefered knowing about these stories earlier on as they would of helped maybe in the structure and flow of your post.
This is where clustering comes in, you fire up your RSS Reader, you click on a feed or maybe a folder river of news, and you read a story on the new Yahoo! Pipes. Under this story is another story on Yahoo! Pipes from another feed…but this feed lives in a folder at the bottom of your subscription list, so you ask yourself what is it (the story from this feed) doing here. Well what it is doing is auto overriding and appearing of its own accord as it’s in your interest, as you wouldn’t of come across this feed/story for at least an hour, or perhaps not at all today, especially if you haven’t got much time to read your RSS Reader.
NOTE: I guess search could cover this in a way, if you find a story you like and want to find similar stories, just do a search limited to unread items, this may do the trick…but then again Google Reader doesn’t have search yet.
Similar ideas to clustering are collating similar stories together based on text analysis, eg. you load in feeds into the RSS Reader, you set up some tags (this is not organising feeds into tags, but moreso just listing areas of interest), and the system will displays stories in these tags.
So this is not clustering similar stories, but moreso collating stories on the same topic…see MyFeedz, and WizAg (actually I think this has both features: clustering and reading stories by machine tag). Also see TagCloud, Zoom Clouds and Personal Bee.
Now you have clustering as your ally, you read a story and similar stories are underneath it, saving you to find them yourself.
NOTE: another idea is for it to cluster items outside of your RSS Reader eg. here is a similar item from a feed you are not subscribed to…this could appear in a separate box, so you could ignore it if you so choose.
Personalisation
Clustering saves you time by collating stories that are the same, but what if this story that you read on Yahoo! Pipes surfaced in your RSS Reader after half an hour of reading. If your time is limited to 15 minutes this is not good enough, as you have just missed out on reading a story that is important to you.
The next ally you need is manifested from your actions, ie. your reading behaviour.
The idea is that when you launch your RSS Reader you will see your river of news ranked with the most important stories according to your attention, and under each story are the clustered items (similar stories), if there are any.
NOTE: maybe you could also choose whether you want to include popularity as a ranking of these stories, this is an influence outside of your reading behaviour..see OPML Sampler : Popular posts within an OPML.
Summary homepage
To augment this even more would be for the Top 10 stories in your RSS Reader at a given time to display on a page with collated stories underneath. This page would be similar to your Google Reader homepage, but this just lists the last couple of latest stories. The page I imagine gives you a brief on the most important stories (plus clustered stories) to you, where you can read them without really having to go into your traditonal RSS Reader 2 pane mode.
More
1. Clustering (mentioned above)
2. Personalisation (mentioned above)
3. Summary homepage (mentioned above)
4. Search
- limit to a folder or a feed, allow you to select what you want to search, even limit to dates
- every search generates its own feed
5. A unique OPML and an OPML for each folder
6. Clip outside items, see here
- although they are not RSS Readers; Ziki, Jaiku, and Top 10 Sources are all different services that let you clip or add content into an automated content stream
7. Collect stories for future blog posts
- I can easily “star” these items, but my starred items are for stuff I want to read later, I want a section for stuff I want to blog about. I could tag these items, but I can’t keep them unread as when I finish reading a folder of feeds I mark all as read. When I view items in this tag, they are all read, I can’t tell which ones I just added (I’d have to mark them unread again, otherwise later on I won’t remember if I’ve blogged about it)…see more.
- I guess I want the ability to create other icons similar to “star”…this won’t fix my issue, but at least I have items that I want to blog about in a separate stream from my starred items.
8. Limit view All items (ie. view all items except these folders)
- reason I want this is that I have one folder that I label “heavy”, this contains feeds that have 10 posts and over a day, I don’t want this in my “All item” river of news.
9. Be able to order folders or feeds
- this way I can have my favourite folder on the top of my subscription list, at the moment I have to do this eg. AAAFavourite feeds
10. You can rename a feed, but you can’t rename a folder.
11. System tray pop up toast notifications
- even configure this to just one folder or feed, or select what you prefer…see more.
Related:
RSS Reading different views
RSS Reading: recommendations
Google Reader and clipping in general

















![[VoiceIndigo Mobilize - Listen to podcasts on your mobile phone]](http://www.voiceindigo.com/ht/images/mobilize_this_disabled.gif)









Very insightful, thanks!
Just to give you more details on how myFeedz works: it’s a combination of clustering and collating. We refer to these stories as “related articles”, because myFeedz also tries to identify the original story (the most important story, that people blogged about).
myFeedz also offers personalization, summary page and search. However, search within OPML folders is not possible, but it’s an interesting suggestion. We found that not so many people are familiar with OPML, so we haven’t focused too much on OPML-related features.
All the best,
Marius
Comment by Marius — February 13, 2007 @ 6:16 pm
Many Web 2.0 Sites Fail to Link to Related Content
For all of their surging traffic, a lot of blogging, social networking, community and other Web 2.0 sites truly fail at keeping people on their site in the most basic way - by pointing people to related content within their
Trackback by Micro Persuasion — February 19, 2007 @ 9:16 pm