The Next Killer App is to Twitter as 1-2-3 was to Visicalc

Friday, April 24, 2009 by Dave Winer.

This post is dedicated to Mitch Kapor, Jonathan Sachs, Dan Fylstra, Dan Bricklin, Bob Frankston, Ben Rosen, Ev, Biz, Jack, Bijan, Fred et al. A lot of people on that list, and I have had the privilege to know most of them, and I've met all of them.  

First a very brief story of what Visicalc was and how it was surpassed by 1-2-3. And please understand this is my version of the story, I'm sure all the principals will have their own versions. 

Visicalc was god. It was the Twitter of the early 80s. It was credited with creating the personal computer boom led by the Apple II. The product was created by the two Dans and Bob. Mitch worked for the company. Everyone said Mitch wrote spaghetti code. The two Dans and Bob didn't take him seriously. They were gods after all and Mitch was a mere mortal. 

Along comes the IBM PC. It's a private thing, but since Visicalc rules, they got early access. The two Dans and Bob and Mitch all saw it. They started to port Visicalc to it, but didn't do anything special. After all, their software is what made the PC boom. The IBM PC would be just another Visicalc machine. Mitch didn't see it that way. So he got together with a brilliant coder, Jonathan Sachs (so much for the spaghetti code!) and got money from a very smart man in NY, Ben Rosen, and started Lotus to make 1-2-3. The two Dans and Bob knew about it, but they didn't take it seriously, because they were gods and Mitch writes spaghetti code. 

But the one thing they didn't figure out about Mitch, that made him such a killer, was that he used the product so he knew what features would be most valuable to other users. Not saying the two Dans and Bob didn't, I'm pretty sure they did, but Mitch was really tuned in and watched users get confused and hung up with Visicalc, so he knew what to focus on for 1-2-3. I don't know for sure, but I bet the Visicalc team didn't really listen to Mitch. 

When the IBM PC came out everyone wanted new software for it, and the Visicalc guys just offered the same old stuff, but Mitch's software had a sexy UI (for the day) and ran like a bat out of hell, used all the memory of the PC, and had a macro language, so everyone bought 1-2-3 and that was the end of Visicalc. 

So somewhere out there is an idea for Twitter that, like 1-2-3, will represent the future, leaving Twitter to own the past. The challenge for brilliant software designers everywhere is to figure out what that is and to do it! 

I've been studying it as long as anyone, I started using Twitter in the summer of 2006, and have been puzzling it out every damned day, waiting for Twitter to give me something new to sink my teeth into, and I'm convinced it's going to come from a Mitch-Jonathan-Ben combo out there, not from the original team. Probably for many of the same reasons Visicalc didn't rise to meet the needs of experienced spreadsheet users.  

Here's my wish list: 

0. It starts as an exact Twitter clone. Command for command. Then see item #2. I get to completely redesign the UI. 

1. I want to start my own Twitter, for free. You host it for me. Anyone can join.  

2. I have to be able to edit the template, fully, so I can make it look like my blog. This will allow designers, for the first time, to tinker with the look and feel of a Twitter. They played a big role in the blogging bootstrap, but have mostly been sidelined by the emergence of Twitter.  

3. I want to map my domain to it, so it's part of scripting.com. 

4. It's gotta be fast! 

5. Lots of prefs that determine who can join, what they can do, various editorial roles. If you used Manila, I want to be able to delegate to managing editors and contributing editors. 

6. Easy hooks into Disqus (and competitors) so each tweet can be the beginning of a conversation. 

7. Plug-ins that hook into the UI so I can add commands to my Twitter, without modifying any source code. 

8. The ability to attach a picture, movie, MP3 or any arbitrary data to a tweet, basically the same power as the RSS 2.0 enclosure element. 

9. Full data portability. I've got to be able to run a script on my desktop every night to get a complete XML-based backup of my community.  

I guess my point is this -- soon there will be enough Twitter users who yearn for something really new, and it seems doubtful that Twitter-the-Company will want to give it to them. With all the new users just getting started, they're going to focus on getting them up the curve. So we're really getting ready for the 1-2-3 of Twitter, the next level of power, so we can build richer and more connected networks. 

See, that's what I really think Twitter is -- a Network Construction Kit for Real People. Sort of a Tinker Toys or Lego for networking. We've gone a long way with a few simple pieces. We need some more stuff to play with. 

It's always like this, in every layer. At first we need training wheels, and a tech company to provide the whole package. Then we get comfortable with the technology and we want to order a la carte, to design our own meal. You can try to contain the users, lots of companies have -- but it never works. If this blog is about anything, it's about that -- documenting the never-ending cycle of tech booms and busts, bright new days, and endless platform wars, starting in 1994. It's all so predictable, you'd think one of the rising stars would figure it out and plan accordingly, but it seems they never do.