The Technology Whiplash Effect on Push
Every so often, a brand new idea comes along in technology. This post is not about one of those ideas. This post is actually about an idea from 1997: Push. You see, push has suffered from the Technology Whiplash Effect for the past 9 years, and I think it's primed for a comeback.
First let me explain the Technology Whiplash Effect effect: I just made it up. It's just a way for me to describe what I have observed when a new idea in computer software comes along. When a new idea comes along, a predictable cycle seems to follow. First, somebody dreams up a fantastic, 
world-changing software idea and scribbles it on a bar napkin. If it's a good idea, people get excited when they think about it. Then it picks up steam, the possibilities are endless, it's the beginning of a revolution, everyone will make millions, etc, etc. It makes the front page of Wired. 
And then every software company in the world is scrambling to band-aid it onto their product so they can issue press releases about it. A few new startups are born around it. Everyone knows 
that the technology promises really cool stuff and they don't want to get left behind, so they try to adopt it before they really know what it's for. And inevitably, a process without a plan is going to fail miserably. And so it does, and then everybody thinks the technology wasn't able to deliver on any of its incredible promises. The people who dreampt it up originally probably would be hesitant to bring that fact up in conversation. It's technology taking off too fast.
Fast-forward five years or so, and somebody is in the process of addressing a real-world problem... but they find that they can't deliver it with the technology toolset they have available to them. So they go back and dust off one of the old ideas, or arrive at it again by re-inventing it. Of course either way they give it a different name, and release it as something new. A few geeks recognize it as a repacking of old ideas, but not many people are listening and even fewer care--the technology actually works and solves problems. The original idea is now solving real-world problems, and all of a sudden it's big news again, under a different name. THIS time, because 
people can actually apply it to something and wrap their heads around it, people start working with it, and addressing real-world problems with it, and it's great, and bubbles inflate around it (again). Oh, and it makes the cover of Wired again.
The cycle seems to have been pretty reliable in the last decade; it's happened to rich Web clients (AJAX is DHTML's second coming), e-commerce (large individual sites have taken over while online malls have died), 
personal web presences (MySpace succeeded where GeoCities failed), and software as a service (online software has been in the news a bit for the past couple of years, long after Pets.com passed on). Now, I predict, the same is going to happen with Push.
The difference this time around is that Push is being used properly, which is to say that it's being used for messaging rather than content delivery. The first time around, everybody tried to push content to users, most (in)famously with Pointcast and Microsoft with its channels built into Internet Explorer. I've heard some people say that RSS and ATOM are equivalent to modern-day Push, but they're wrong. Those technologies simply give people a way to find out what's out there and pick it up if they want it. An RSS feed doesn't push content to users any more than a vending machine pushes a pack of gum to a person. It's a polling (pulling?) mechanism.
Real bonafide Push, on the other hand, actively goes out and tries to deliver something to a user, and it's on the verge of a comeback. This time what we'll be pushing isn't content itself, but update notifications and pointers to the actual content. As it always is, the best solution is a combination/compromise of two ideas--PUSH the notification telling somebody that there's new content, but let them decide if they want to go PULL it to pick it up. The bandwidth problems that killed Pointcast would not have been an issue if they would have used the technology to push notifications instead of actual content. M. David Peterson, Russ Miles, and several others have done a great job creating a specification called LLUP (PULL spelled backwards) around these principles.
The possibilities once this ball really gets rolling are pretty wild. See my post about sharing rich data if you're interested in the kinds of things I personally am trying to do but can't with current technology.
Mark Cuban had a really great blog post the other day about the Internet in general, and he summed it up well by sayinig "There is nothing 'oh my god' unique that has happened on the net in forever". He's right, it's probably why I find Tech.Meme so boring these days. I think it's about time for something new (or old, as the case may be). And I personally am at a place where I need to revisit push in order to do what I need to do. Thankfully, other people are seeing the same things that I am and have already started working on doing it right the second time around. So although this sounds really weird, and I'm sure at least some people will laugh out loud at this, I'm working on new projects with push as a core tool.
