Google Wave
Vidar Hokstad has a really good summary of Google Wave, focusing on some of the deeper aspects that were kind of overshadowed by the product demo. I spent a good chunk of the weekend poring through all the Wave documentation, and I’m feeling pretty excited about it. One of Vidar’s main points is worth repeating: Wave looks great not because it’s completely new, but because it’s a distillation of a lot of existing ideas into one concise, powerful, and accessible protocol.
Just to touch on a few items that Vidar doesn’t hit directly:
If you spend any time on IRC, you’ve seen the bots (for instance, Haskell’s beloved Lambdabot). Wave stands a shot at bringing these amazing little guys into the mainstream. I think it’s highly possible that robots will emerge as the major business opportunity that springs out of this platform. Reading around, the current favorite right now is corporate customization and installation, which will no doubt have huge revenues, but those revenues are going to be offset by higher costs. Robots will be as cheap and easy to get up and running as a website is today, and many bot services will benefit from the network effect.
Adoption is, of course, a concern. How many people are going to jump on board, and how quickly? Something the demo makes clear, though, is that there are many hooks into Wave in the world as it exists right now. Blogs. Twitter. I’m sure some lucky developer with sandbox access is already trying to tie it into Facebook. The point is that regular people will start using it without having to make the mental leap of deciding to try something new, just by coming across an embedded instance in a site they’re already comfortable using. The web won’t have to shift to Wave wholesale in order for it to have value, it’s going to be able to grow organically.
Finally, the instant Wave is released, it will already be a commodity. Depending on the level of real interest, there might be something of a goldrush to pick up the open-source server software and package it up for various use cases, but really, everyone’s going to be on more or less the same footing right away. I might be totally wrong on this, but is this the first potentially major protocol that’s going to start its life this way? It will be interesting to see how many 3rd parties try to build their own servers from scratch, though I’m sure there’s already someone holed up with the spec going nuts with it now.