Text Patterns - by Alan Jacobs

Tuesday, December 9, 2008

the short goodbye

The coolest tool for personal organization I have ever seen is called Stikkit, but if that link works for you it probably won't for long. More about that in a moment.

A good and fairly detailed overview of Stikkit’s features may be found here, but in brief, it works like this: Stikkit gives you a field into which to enter text, and uses an extremely smart natural-language parser to figure out what kind of information you have entered. Just type in, or paste in, someone’s name and contact information and Stikkit recognizes phone numbers as phone numbers, email addresses as email addresses, and so on. (Incidentally, a similar intelligence was featured about a decade ago ago in Simson Garfinkel’s address book SBook, but Garfinkel abandoned that project. A real shame.) Tell Stikkit “remind me to go to lunch with David next Saturday at noon” and it puts that item on your calendar and sends you a reminder. Your contacts have short names (davidw, say) so that you can type “lunch with davidw” and Stikkit includes a link to that person’s contact information. Stikkit is also unusually attractive, especially for a web app.

I started using Stikkit right at the beginning and thought it was the most awesome thing I had ever seen. Yes, the parser sometimes made mistakes, but I could correct those, and it was at least 90% accurate. I expected that further development would make it even better, and I became a heavy user, offering frequent feedback to the Stikkit team, who seemed to appreciate it. The service was free, but I expected that at some point it would cost me something, and I was ready to pay.

But then development slowed, and eventually stopped. The Stikkit team stopped answering my emails, and stopped answering questions on the forums. Then they announced that they were going to be using the Stikkit engine to undergird a different project they called Sandy, which, they said, would be an even better way to do what Stikit does.

Well, I took a look at Sandy and didn’t care for her at all. But even if I had, my suspicions were raised. If the team — the company, by the way, is called values of n and is headed up by the gifted programmer Rael Dornfest — could so quickly and silently abandon Stikkit, might they not do the same with Sandy? I immediately decided that I couldn't trust them with all my data, so I exported it as best I could and found other solutions.

Good thing, too, because values of n is shutting down, and both Stikkit and Sandy are going offline on December 19. Read the comments of the post I just linked to, or the comments here and you’ll get just a taste of the anger that’s being directed at Dornnfest, who seems to be a genuinely nice guy who never expected such a response.

I don't feel I can blame people for being angry. They have invested tons of time, if not money, in a system that is now going to disappear in a puff of smoke, another Web 2.0 casualty. And it seems odd that Dornfest, as far as anyone call tell, never even sought a revenue model: many people are saying that they would be more than happy to pay for Stikkit or Sandy if Dornfest would just give them the chance. But nope, it’s all going to evaporate into the ether. Very sad, and puzzling. But no one has a legal (or a moral?) obligation to offer an online service any longer than they want to — especially if that service is free.

Dornfest, interestingly, is going to work for Twitter, and will be bringing along some of the technologies he has developed. But in what form? And will anything of the brilliant Stikkit model of free-form data entry survive? And how concerned should we users be that Twitter has no revenue model either?

What are the lessons of all this? I can imagine a few, but I’m wondering what you readers think.