RSS, Human Filters And Real Time Streams

It has been a few months since the “RSS IS DEAD!” meme popped up, but the release of some new tools has sparked conversation again. I am ignoring the people who just say that they use Twitter to get their news, because you can easily miss too much no matter how diligent you watch your stream. Marshall Kirkpatrick loves the idea that other tech writers are moving away from RSS and towards Twitter or Techmeme:

I am having a hard time believing that he (Sam Diaz at ZDNet) and so many others say they no longer even bother to read feeds. Twitter, Facebook and aggregators like Techmeme or Google News suffice for Sam, he says. He’s far from alone.

Robert Scoble provides a different perspective, mainly focusing on the fact that people like us (Marshall, Louis Gray or me) are not normal. His focus is mostly on breaking news, which he says normally ends up on Twitter:

It’s very rare that I see news in Facebook or FriendFeed that I don’t see in Twitter first. It’s even rarer that I see news in Google Reader. Why? RSS isn’t real time.

Robert does state that you need to follow a select group of people with your interests in order to really follow news. His feeling is that we need better filters. In that respect, I completely agree, but I wanted to talk about the main points people seem to be talking about, RSS, human filters and real time.

First, RSS is a data specification, not a protocol and not an application. RSS is that xml format that defines a “syndication” format for your data. In many cases it is an xml translation of your blog post that an RSS Reader will use. RSS does not deal with timing, RSS Readers do. The ping protocol and things like Weblogs servers definitely deal with timing, and that is where RSS and real time get confused. There is nothing stopping people from hooking up their wordpress blog feed to a PubSubHubbub server and that has already been done by many people. However, no matter how fast the data is pushed around, RSS readers are not real time and typically poll servers on some scheduled basis. Google Reader is slowly moving towards PubSubHubbub, and are posting shares through those real time servers.

However, getting real time access to posts in Google Reader does not matter. What matters is what you are reading and whether it is interesting to you. Aggregators like Techmeme or even Topsy tend to find the most popular, most talked about or most shared articles. Typically, this means they also have the most readers. Granted, sites like TechCrunch, ReadWriteWeb and Mashable do post some interesting articles, but they tend to be more focused on breaking news. They will be highlighted on all of the aggregators, but are they interesting to you?

This is where Scoble’s thoughts about human filters becomes important. How do you find the appropriate human filters for what you find interesting? Twitter can be helpful in finding people who share things on Twitter, but that is not the full picture. For most people, meaning not early adopters, you would really need to look at Twitter, Facebook and maybe Google Reader. Google Reader is not really mainstream, but it is an RSS reader and there will be very little noise to deal with. I have previously stated that I could be your filter for much of your technology information. But, how do you find that out? How do you know that Michael Fruchter shares tons of good marketing information?

Disappointingly, I do not have the answer. Some people are using tools like Feedly and LazyFeed, and maybe they will figure out the discovery problem. Maybe Google Reader’s new social push will use your connections to aid discovery. Or maybe you are happy reading Techmeme and do not care about breaking news. The main thing is that we are all trying to find interesting information, and we are having a difficult time doing that.

21 thoughts on “RSS, Human Filters And Real Time Streams

  1. Jeff,

    That is part of the point. There are some people that want to know within a minute or two that something was posted. For others, sometimes an hour is OK.

    But, was it interesting to you? I hope it was 🙂


  2. For what it’s worth, you’re a wonderful filter as are Mike and Robert too. However, even with you guys it’s still hard to find what I’m looking for. I read an interesting post by Mark Cuban about webhooks and PubSubHubbub yesterday; Even though it’s not important to get real-time feeds, it still surprises me that it’s taken so long to get around to pushing rather than polling for information. Unless you’re making a living trading in the markets, real-time data isn’t necessary. I’m sure there are other fields where it’s important as well, but they usually all have their own services which provide for their needs. It’s nice to come across a breaking story and share it with others before anyone else has had a chance to read it, but you’re right, “What matters is what you are reading and whether it is interesting to you”. Well said!


  3. “However, getting real time access to posts in Google Reader does not matter. What matters is what you are reading and whether it is interesting to you.”

    You’ve brought to light a great nuance surrounding the movement to real-time access to info and how filtering fits into the equation. The above lines say it all for me.


  4. Michael,

    Real time has very specific applications. Some are things that you mention, like trading markets, news is another area but mostly from the journalism side. News consumers do not typically need to know in real time.

    That being said, we would all like better information faster. That means if we get real time we want way better filters than we have now.


  5. pixites

    You are correct, I have not tried lazyfeed yet. It is on my list of things to do, but I need to find some time.


  6. autom,

    I am glad you found something that fits your feelings. As I mentioned in my response to Michael, real time info needs better filtering. Even Scoble basically admits this in his post.


  7. My proposed solution involves building software tools that aid in learning about user likes/dislikes by dissecting our social broadcasts (blogs, twitter status, public facebook status, friendfeed, etc) using the best semantic tools available (I use Zemanta at the moment).

    From a database of tags with weights, and frequency of occurrence real time and offline search can be done to find potentially relevant information for me on a periodic basis. In addition relevant advertising can be put in a container that I have control over (scary huh?). When I want customized, and relevant ads I’ll open up my ads tab and see what businesses are interested in my attention.


  8. Michael,

    Recommendation engines are possibly helpful for finding who shares similar content, but have been lacking for blog discovery. I think it is because there are better data points with sharing behavior than whether you would like a specific blog.


  9. Mark,

    Outside of the semantic aspect of your proposal, that is very similar to what SocialMedian does. It uses weighting of your interests and sources to determine what seems more relevant to you. Though they did not get into advertising.


  10. As far as semantic/suggestion engines go, it’s too early to know how useful they will become. I’ve tested several recently and I’m convinced they will play a significant role in our lives going forward. They are so good at figuring out what I like it’s almost frightening at times. If they can deliver equally good results in comparison with anything else out there, then it’s their simplicity of use which will ultimately lead to their success. BTW, I signed up for my6sense today, and it sounds a little like what Mark is doing, minus the creative advertising. BTW, which Scoble piece are you referring too? I must have missed it.


  11. Hi there

    Very interesting topic if you ask me. I came across Lazyfeed (and indirectly this post) as it provided a reference to my own blog today, so in a bizarre way I was ‘pulled’ here.

    I’m interested to try a few of the links mentioned in the post above.

    I also recently posted a blog on this topic, titled “Ensembli – a solution to information overload?”

    I’d be interested in any comments. When I posted, I’m not sure that I was aware that there are other (or so many other) options in this arena – I nevertheless find Ensembli very good, and the team there has lots of ideas how to develop it further.



Comments are closed.