An Emergency Broadcast System for Twitter might look like this. ![permalink](http://scripting.com/images/2001/09/20/sharpPermaLink3.gif)
![](http://scripting.com/images/2011/11/29/blank.gif)
![permalink](http://scripting.com/images/2001/09/20/sharpPermaLink3.gif)
When everything is working at Twitter, all the clients communicate through Twitter. A client is an app like Tweetdeck, UberTwitter, Seesmic, Brizzly, Tweedroid, etc. ![permalink](http://scripting.com/images/2001/09/20/sharpPermaLink3.gif)
However, when Twitter is down, each client sends your tweets to a "safe place" and the clients your friends use hear about your status updates from there. ![permalink](http://scripting.com/images/2001/09/20/sharpPermaLink3.gif)
There's still a lot of disruption, but some messages get through. Over time the EBS will get better, as we learn better how to make the best of a Twitter outage. ![permalink](http://scripting.com/images/2001/09/20/sharpPermaLink3.gif)
Reality: We still need to communicate even when Twitter is down. ![permalink](http://scripting.com/images/2001/09/20/sharpPermaLink3.gif)
Bottom-line: The client vendors are key. ![permalink](http://scripting.com/images/2001/09/20/sharpPermaLink3.gif)
5/20/10: "The secret for the client guys is that instead of investing in individual platforms owned by huge corporations, they must swallow hard and invest in working with each other." ![permalink](http://scripting.com/images/2001/09/20/sharpPermaLink3.gif)