Skedsheet Blog

I’m wondering at what point should we let anyone outside Moraware start trying out Skedsheet?

Conventional wisdom in the internet software/web 2.0 world says that you should release early & often.  I certainly agree with this concept, and I’ve even given that advice to others before. But when you’re trying to apply it to your own situation, things seem more complicated.

Skedsheet is in the very early stage of development. We’re not working on the beta yet. We don’t even have an alpha build. We’re currently working on the second prototype!

The first prototype from this winter was basically designed like this:
Ted (aka product manager): “It’s called Skedsheet. It’s a spreadsheet combined with a calendar. Go.”
Derek (aka developer): “Uhh…OK”
So it wasn’t my best management moment. But we got the first prototype done, and it still looked as interesting on the screen as we thought it would when we first dreamt it up.

We had to step away from it for a bit to finish up JobTracker v3.0. As we wrapped that up, it was time to take a second pass at Skedsheet. This time, Derek & I got together to hash out as many ambiguities as we could, and decide what we would do for the next milestone: the second prototype. This meant a lot of reworking how Skedsheet behaved, and making many decisions about what functionality we wanted to provide.

So now I’m looking at a half-finished second prototype on my screen, and it’s pretty cool. Much better usability than the first pass, but there’s tons of functionality yet to be done. And it’s butt-ugly.

There’s a reason I’m calling these prototypes, rather than alpha builds. I’m not yet convinced we have the right user-interaction or the right client-side technology choice. When building a brand new product, there are a daunting number of choices to be made, and picking just the right combination is quite a challenge. We may end up completely rewriting major portions of the application as we go through the development process.

So when should we let others try it out?

• As soon as it does anything at all? (i.e. right now)
• As soon it has a critical set of features completed? (what’s really critical?)
• Only after we feel confident that it solves the problem well?

There are a few risks with releasing too early:

1. It takes longer to do development because we have to worry about existing data and users that we don’t want to break.
2. We put out something useless because it’s missing too many important features. So we lose the opportunity to get people interested the first time they try it.
3. We are still actively working on JobTracker, so Skedsheet is not going to be moving forward quickly at all times. So anyone using it early may be stuck with problems for a while before we fix them.