inessential by Brent Simmons

One of My Mistakes

I don’t think of my app Glassboard as a competitor of Path. Though they’re both superfically about sharing statuses, locations, pictures, and so on, they have different purposes. And Path is a beautiful app that I admire very much.

Nevertheless I admit to feeling a bit of schadenfreude at first when it was discovered that Path was uploading contacts info to their servers.

I’m human, in other words, and humans often react in ways beneath them — until they try a little harder.

My next feeling was sympathy. I’ve made plenty of mistakes (and I’ll tell you about one in a minute), and other people have surely enjoyed my errors from time to time.

Here’s the thing about uploading contacts without notice or permission: it’s wrong, and the Path folks know that. But it’s also true that launching a new social network is difficult. There’s a great deal of friction. You want to make it as easy as possible for people to connect to their friends in the system and to add new friends — and people don’t want to type in a bunch of email addresses.

A new social network needs to make this as easy as possible. It should seem like magic; it should “just work.” And every app maker should have user experience as their top concern: they don’t want people to struggle to make the app useful.

It’s easy to see where a focus on user experience, on reducing friction, could lead to the decision to upload your contacts to their system. It’s still wrong, but you can understand it — because you understand wanting to make an app with a great user experience that people like. Everyone tells you that that’s the most important thing.

Here’s the question that interests me: did uploading contacts allow them to include functionality that allowed Path to grow faster than it would have otherwise? Might it even have made the difference between a successful product and an unsuccessful product?

There’s no way to answer that question.

But I’ve long had a similar question about NetNewsWire. Back in 2002 and 2003, when NetNewsWire would read a feed, it would send the URL of NetNewsWire’s product page as the referer.

That was wrong. (It’s also what other readers did, so I justified it as the then-current best practices.) It wasn’t a privacy violation — it wasn’t terribly wrong — but it was still wrong because it was misusing the referer. (The user-agent is where that info was supposed to go.)

The effect, though, was that bloggers who looked at their stats would see the NetNewsWire product page high in their referers list, and they’d check it out, and then more people would see the app and use it.

Did this make a big difference in NetNewsWire’s adoption in the early days? Did it even make the difference between a successful and unsuccessful product? I like to think it didn’t make much difference, that I could have — and should have — not done that. But there’s no way to know. And now, ten years later, I wish I hadn’t done that, even though it was the convention.

Genesis Planet

The lesson: don’t use protomatter in the initial matrix. Things will blow up.

Even after this, I trust the Path folks to deal with this and do the right things. (And I wish them well. It’s a fantastic app.)