SuggestRSS: Super Simple New Feed Recommendations
Let's say you like RSS feeds. Let's say you're looking for some more good ones to subscribe to so you don't miss good stuff. There are a number of ways to accomplish this, but this week New York based developer Jonathan Christopher built a nice simple new one. Give his app SuggestRSS just a few minutes of your time and it will give you some cool new feeds to subscribe to.
Years after the untimely demise of Share Your OPML and available now, unlike the super feature rich but frozen in private beta service Toluu, Jonathan Christopher's SuggestRSS is easily worth the minimal effort it will take you to try out.
To use SuggestRSS, export your reading list from your feed reader and you'll get an OPML file on your desktop. Upload that file to the SuggestRSS website and it will do a simple comparison. People who like what you like also like these things you don't have. Simple as that. Amit Agarwal found and blogged about it first, among people we know.
You'll get some duplicates with the system, the sample size is small, it's dominated by tech feeds and not good for much yet with other niches. But you can come back to the URL of your recommendations later and find new suggestions!
That's it. There's nothing more to it. Now if you'll excuse me, I can't believe I had never thought to subscribe to the feed for Boston.com's Big Picture. It's really incredible photography and I appreciate SuggestRSS prodding me to subscribe.
Share
Comments
Subscribe to comments for this post OR Subscribe to comments for all ReadWriteWeb posts
How about a FilterRSS? A simpler way to READ feeds? :)
Posted by: Rui Pereira
|
January 21, 2009 5:28 PM
Wow this is sort of like our site, except we do this for local services. Even the names are very similar :)
I wonder how accurate the suggestions can be given that for our site we needed 100's of thousands of reviews to get really accurate predictions. Another interesting concept is that this treats all items in your OPML the same. I wonder how much more accurate the results would be if after importing your OPML you could rate the feeds.
That said - I really like the simplicity of the idea and the fact that you can get personalized suggestions with a single import.
I think this service seems quite biased...
I only have 1 apple based feed in over 200 feeds in my OPML and most of the feeds that were "suggested" to me were Apple feeds, like MacRumours and so on...
Also I got suggested the Boston Globe's Big Picture, even though I don't have any kind of news feeds like CNN ...
Would a blog give kick backs for adding subscribers?
I don't know... but Google Feed Suggestions were more matching to my own feeds.
What I really need though is a service that goes through my opml and deletes feeds based on how much(little) time I spend reading items from a particular feed. Seriously, who really needs more new sources with the # of dupes, top 10 blah blah blah lists and other crapification rampant in RSS these days.
I just tried it. I pushed through my 212 feeds and got a bunch of suggestions.
The highest "probability" was 35% that I would like it.
Well, I went through the top ten suggestions and I only found one that I liked.
Either this thing sucks or I currently subscribe to every good blog known to man.
Please tell me it's the former.
anyway to extract the opml file from pageflakes or other aggregating sites?
i like the idea of analyzing the time spent on each rss and ranking. an old idea i had awhile back:
http://iancamarillo3.blogspot.com/2007/03/this-website-tracks-your-web-usage-for.html
also, is "FilterRSS" "www.feedrinse.com" what is the URL for "FilterRSS"?
thanks
Great Service. Interesting that a lot of people get suggested Boston.com's Big Picture Feed.
Would be nice to get RSS suggestions filtered through categories.
cheers,
Herbert
Did people read the text? It clearly says the sample size is currently small. The more that is added to it, the wider the spectrum of returns you'll eventually get.
As for getting suggestions that you will definitely like, that's like going to a buffet and disliking it because there were carrots in there, and, you don't like carrots.
I say, let this thing flourish and over time the sample-size will grow, thus the law of probability will deliver more targeted and more 'likeable' feeds to feed on. Yum.
This is great, thanks. Always looking for the newest tools like this, but must say, the my problem is becoming too many feeds. Managing them is taking a great amount of time!
Thanks so much for the writeup! I've spent some time optimizing and revising the suggestion algorithm and speed should be improved while duplicates should be reduced. Continuing, I'd like to stress the fact that SuggestRSS was designed with one thing in mind: simplicity. My aim was to set up a service to which you can post your OPML and simply get a list of suggested feeds to check out. The algorithm, by nature, is indeed biased. The suggestions are provided based on the popularity of the feeds in question. Finding any long tail results aren't a part of the scope quite yet, but who knows what the future will bring. Thanks to everyone for taking a few minutes to submit your OPML, the more feeds submitted, but better results will be.
If you take this kind of service to its logical conclusion, surely you will have a situation where (a) everything that is popular was already popular, and (b) everyone likes exactly the same thing.
I'd like you to discuss the underlying impacts of (useful) popularity-based tools like this and your beloved PostRank. I'm a little uneasy about this obsession with popularity and crowd-sourcing. Crowds can be wise but also crazy and dumb.
Rollo - popularity is certainly no substitute for quality, but it's a useful if crude pointing tool that can be measured. We've got other methods for discovering quality things online other than popularity, we just can't talk about those methods for competitive reasons :) Popularity based tools are useful as well, I believe.
How well does this work compared to google-reader?