On Saturday I wrote a piece about Windows NT, the operating system Microsoft created to run alongside the original Windows, as its eventual successor.

In a similar way, that's what Twitter needs. New technology, running off on the side, for pioneering users, the kind of people who were attracted to Twitter in the first place, the people who created the culture of Twitter.

Let's get the band back together, and this time, let's use what we learned the first time around to do it better.

New technology for Twitter

Here's what I would like to see in Twitter NT.

1. Easy programmability via a scripting language (JavaScript of course) that works directly with Twitter's API, in the browser. This is within reach now.

2. A more powerful data model. Let us attach any JSON structure to a tweet. There's plenty of prior art here going back to the original Mac toolbox, which allowed developers to add a refcon to a window. From that all kinds of amazing things blossomed.

3. Allow us to define in-browser callbacks that determine how a tweet is displayed, using the JSON data to either guide the rendering, or to provide the data being rendered. Instead of trying to force uniformity, let a thousand flowers bloom.

4. This new-technology-Twitter is walled-off from the celebrity version -- the one that entertainers, politicians, sports heroes and their fans use. Our tweets don't exist in their world, and vice versa. What we get is a solid back-end based on Twitter's cloud. And a license to explore and hopefully create the future.

5. Give it a few years to gestate, new ideas will flow from the place the hippies play to the news system of the future. This is where we've been aching to go with an Internet scale notification system.

In 2011, Twitter took a turn away from being a platform. This was, imho, the big mistake. Now the right thing to do is become the realtime platform that still does not exist on the Internet. Twitter is in the best position to provide it, for now. That opportunity won't exist forever.