Search Engine Land schrijft over een bug die al 2 jaar in MSN bot zit. Dat is de spider van achtereenvolgens MSN Search, Microsoft Search, Live Search, potentially Kumo en nu Bing. Webmasters krijgen dankzij MSN bot vreemde resultaten in hun logs en Microsoft is er al 2 jaar van op de hoogte. Reactie van Microsoft: stel de vraag even via ons forum.
En dat moet Google verslaan?
I see the advantages of microblogging. Really, I do. But I hate Twitter. I hate it when a technique or market is almost completely owned by one company. That’s my main problem with Google when it comes to search, Adobe when it comes to Flash and Microsoft when in comes to operating systems.
When looking into the specifications of pubsubhubbub (please change this name guys, seriously) I suddenly got the idea that this was a missing link in the set of protocols that could replace Twitter. It did wonders for Google Reader, maybe it can do more.
If we want microblogging, and the real-time web (or the 2010 web as some people like to call it) as a whole to move forward, we need something open and distributed. What happens when Twitter is down? Exactly. It’s nice and easy to have a central point of communication, but when it’s down, everything stops.
We just need all big players in the blogging and social networking field to join in an effort to get rid of Twitter for once and for all.
I think there are 4 things about Twitter that could be replaced if we all really wanted to:
Let’s call the new techique Cactus (Checking and Contributing to User Status), because we really need a new buzzword and because it’s easier to write this post with a name for this new technique. :-)
The microblog
This is the easy part. We already have more than enough microblogs. Every social network, every blogging platform and every website could (in part) be a Cactus microblog, as long as the supply a Atom (or RSS) feed with a link to a pubsubhubbub hub server.
The aggregator
Everyone can be a status aggregator. Let’s say Google started a Cactus aggregator at cactus.google.com and let’s say I’m a user there. If I wanted to follow for example matt.wordpress.com I would click a ‘follow’ button on his blog, which directs me to my status aggregator. Google Cactus then registers to the stream coming from wordpress.com’s feed hubs and receives instant updates from Matt’s microblog. All Google has to do is cache this stream to it’s own database and supply me with an overview of all updates and subscriptions. It’s like a miniature Google Reader.
The aggregator should also be able to post to the microblog. This is where Oauth and XML-RPC come to mind.
The communication tool
This part is a little tricky. Twitter users have come accustomed to the use of @username, but since we don’t really have a username in the Cactus system, we don’t know to whom the post is addressed. We could however let the aggregator’s user assign a short name to each subscription. In the case of matt.wordpress.com it would simple be @matt. The aggregator and microblog could then render @matt with a link to matt.wordpress.com.
With RSS and Atom we’re not limited to 140 characters so I think this limit (if any limit at all) should count for the amount of characters when rendered in HTML. This way we could get rid of bit.ly and the likes as well, they are an unwanted and unneeded layer in the real time web.
The aggregator should monitor all feed hubs for links to microblogs. This way we know if there’s a post adressed to a certain user.
The discovery engine
This is a main advantage of the distributed Cactus system. Everyone could start a search engine that doesn’t rely on Twitter’s API. This means Google and Bing could both eat a big chunk out of this pie and I think that’s a good thing. Twitter’s search is pathetic and I don’t think it will ever be a search engine I will use on a daily basis.
Conclusion
I’m not really inventing something new here. I’m just using some existing standards, like RSS, Atom, pubsubhubbub, Oath and XML-RPC and applying them in such a way that it could replace Twitter, leaving us with a better, more open and distributed alternative.
Any comments?
