Three Predictions

Ok folks, here we go. I've put the word futurism is in the title of the blog and its near the end of the calendar year, so I think it's fair that I do a little bit of wild prediction about the coming year (and some change.) Here goes:

Open technology will increasingly be embedded into "closed technology"

The astute among you will say, "but that is already the case:" The Motorola Razor cell phone that my grandmother has runs the Linux Kernel (I think.) And there's the TiVo, lets not forget the TiVo. Or, for that matter the fact that Google has--for internal use, of course--a highly modified branch of the Linux Kernel that will never see the light of day.

That's old news, and in a lot of ways reflects the some of the intended and unintended business models that naturally exist around Open Source.

I'm not so much, in this case talking about "openness" as a property of code, but rather openness as a property of technology, and referring to long running efforts like XMPP and OpenID. These technologies exist outside of the continuum of free and proprietary code but promote the cyborg functioning of networks in an transparent and networked way.

XMPP says if you want to do real time communication, here's the infrastructure in which to do it, and we've worked through all the interactions so that if you want to interact with a loose federation (like a "web") of other users and servers, here's how.

OpenID solves the "single sign on" problem by creating an infrastructure for developers to be able to say "If you're authenticated to a third party site, and you tell me that authenticating to that third party site is good enough to verify your identity, then it's good enough for us." Which makes it possible to preserve consistent identity between sites, it means you only have to pass credentials to one site, and I find the user experience to be better as well.

In any case, we've seen both of these technologies become swallowed up into closed technologies more and more. Google Wave and Google Talk use a lot of XMPP, and most people don't know this unless their huge geeks (compared to the norm.) Similarly, even though it's incredibly easy to run and delegate OpenIDs through third parties, the main way that people sign into OpenID sites is with their Flickr or Google accounts.

I'm not saying that either of these things are bad, but I think we're going to see a whole lot more of this.

A major player on the content industry will release a digital subscription plan

I think, perhaps the most viable method for "big content" to survive in the next year or so, will be to make content accessible as part of a subscription model. Pay 10 to 20 dollars a month and have access to some set quantity of stuff. Turn it back in, and they give your more bits. Someone's going to do this: Amazon, Apple, Comcast, etc.

It's definitely a hold over from the paper days when content was more scarce. But it gets us away from this crazy idea that we own the stuff we downaload with DRM, it makes content accessible, and it probably allows the of devices to shoot down (to nominal amounts). While it probably isn't perfect, its probably sustainable, and it is a step in the right direction.

Virtualization technology will filter down to the desktop.

We have seen tools like VirtualBox and various commercial products become increasingly prevalent in the past couple of years, to decrease the impact of operating system bound compatibility issues. This is a good thing, but I think that it's going to go way further, and we'll start to see this technology show up on desktops in really significant ways. I don't think desktop computing is in need of the same kinds of massive parallelism that we need on servers, but I think we'll see a couple of other tertiary applications of this technology.

First, I think hypervisors will abstract hardware interfaces away from operating systems. No matter what kind of hardware you have or what it's native method of communication is, the operating system will be able to interact with it in a uniform manner.

Second, there are a number of running image manipulation functions that think operating system developers might be able to take advantage of: first the ability to pause, restart, and snapshot the execution stat of a running virtual machine has a lot of benefit. A rolling snapshot of execution state makes suspending laptops much easier, it makes consistent desktop power is less crucial. And so forth.

Finally, system maintenance is much easier. We loose installation processes: rather than getting an executable that explodes over our file system and installs an operating system, we just get a bootable image. We can more easily roll back to known good states.

Not to mention the fact that it creates a lot of economic benefits. You don't need IT departments maintaining desktops, you just have a guy making desktop images and deploying them. Creating niche operating system images and builds is a valuable service. Hardware vendors and operating system vendors get more control over their services.

There are disadvantages: very slight performance hits, hyepervisor technology that isn't quite there yet, increased disk requirements. But soon.

Soon indeed.

comments powered by Disqus