Something one of my friends said got me thinking. He was complaining about the iPad not having any communication ports on it, like USB, or serial.
Yea, at first I thought it was crazy too, but in consideration I started to get the bigger picture.
The classic view of a computer was this thing that exists all together in one place, that you interact with it to calculate stuff. The first machines were massive but we lived through an age where they became smaller and smaller. With each downward shift in size, the machines got into a close proximity to our home, our lives and too being invisible. They became ubiquitous.
But still, while the machines were growing smaller and smaller they were still all contained in one place. A machine, and its peripherals were still a single collected entity, all interacting nicely with each other.
At the same time as things started getting smaller, they also started getting more distributed. Sun's "the network is the computer" slogan started a range of efforts to make things distributed, parallel or both.
The computer, went from being a single discrete entity, gradually into a larger number of parts, distributed over larger and larger areas. Culminating with the Internet spanning the world, which is an amazing feat.
But still, even as the 'computing' became distributed, the basic building block was the single 'computer'.
Now, along comes wifi and it gives us this really cool ability to create invisible, transparent networks. The interconnections have been great, but so far limited. We just replace the most obvious wires on the floor.
But what we can do now is start using wifi networks as the back-bone for individual computers themselves. That is, you no long buy a machine for home, instead you just buy the individual pieces and assemble them into you house as needed. Plug and play will become "proximity and play" where just bringing in a new component and giving it the wifi password will be enough to fully configure it into its new environment. It will be easy to extend the capabilities of your house, gradually as you keep buying more gadgets.
The iPad becomes the screen and keyboard, but you could have a separate machine to hold the disk-space, another as backup and a completely independent printer. I've heard of people connecting speakers to their net, and one would like to connect bigger displays like TVs as well. This could also extend to all of the appliances, and even to various types of storage too, like closets.
Basically, the house becomes the computer, the wifi the bus between the devices, and all of the components just connect to everything. You're already living in your next machine ...
Now if we mix this further with the growing trend towards 'natural user interfaces' (NUI), we should start to see more and more intuitive, natural methods of interfacing with these extended machines.
Talking, moving, interacting with objects in the real world. All of these can be tracked, analyzed and enhanced by our equipment. Some day you mother or wife won't have to nag you about tossing you recently worn clothes into laundry hamper, it will be complete automated instead.
The big thing about natural interfaces to me is that we previously had to go to the computer to interact with it. We were on the computer's turf, playing according yo he computer's rules. Now these new interfaces are allowing the computer to come to us instead. That's a big change.
As always with technology, though, while the hardware is going through a number of really exciting new changes and growth, software has been lagging behind, restricting our usage.
The last few decades have seen more code and more data get created, but while they are prettier, computers are way more frustrating to use than ever. We've started a trend towards stupidly complex systems, that are just horrifically painful to work with. Every feature under the sun is packed badly, into some weird or awkward interface, and keep in place by its sheer size. Big, ugly and monolithic.
Each new generation of programmers simply rebuilds the failures from the last generation, but with few real improvements and many many dis-improvements. There will little learned from history.
As an industry, software for decades has been about creating new fads and flaky technology, but not about fixing or improving the essentials.
Even the current mantra now about things like "not creating new frameworks" shows how tunnelled we've become, since all of the existing frameworks are awkward to use, don't solve the core technological problems and all basically use variations on of the same design. One that clearly isn't working.
You'd think we'll be be out exploring new interesting alternatives given our current technologies, but that approach is frowned upon. The current fad is to be too focused on re-writing that domain application (for the third time), so as not to be able to spend any time fixing any of the real low-level core issues. Leave em for someone else.
But then software development has always been driven by fanatics to some degree, although lately the trend is getting worse. Most of what is being offered up as solutions are pretty little band-aids trying to cover up gaping wounds.
I'm literally more embarrassed year after year about the failing state of the industry and it is only get worse. I rarely tell people what my main profession is anymore (I'm just sick of the complaints), more often I'll talk about photography or writing.
Still, someday, software too will grow out of its current malaise. With all of this new hardware we'll need to tackle two very complex problems: a) making distributed things simple and reliable, and b) upgrading to new versions of software.
Both problems don't seem difficult, but neither works right now.
Our paradigms for building software are mostly based around mammoth, single deterministic architectures. Our current technologies are intricate, overly-complex and are easy to mis-use. We've yet to find an abstraction that makes distributed computing trivial.
And as long as it is non-trivial, it will always contain a significant number of bugs, and be difficult to trust. This problem will only increase as the hardware becomes more and more distributed.
Our models of software development are still based on big-design-upfront projects (even if we do release them more often in less completed versions), and thus we have no idea how to release a multitude of small, ever-changing pieces without accruing so much technical debt that the system is mostly unusable.
Common approaches towards backward compatibility and version upgrades are very dependent on significant human effort and thus are costly and frequently prone to failure.
If machines get more distributed, in smaller parts, that also means the upgrade problems that we are having trouble with now, are going to get exponentially worse in the future. More pieces, means more dependency failures.
The hardware will quickly become an interoperability nightmare, a sea of strange undesirable errors and odd behaviors. It is bad now, and it will get even uglier.
If not, it will only be because we'll just stop upgrading the code, and many of the new advances in hardware will get lost for generations. Either way it will be software problems holding us back.
So it seems that the future is going to provide us with more ubiquitous computing, where machines are getting smaller, more specialized, less feature-based and more embedded into their environment. Once software breaks free of it's current dark age, we'll be quietly surrounded by a vast array of different machines all interacting with each other, in other to help us managed our lives and our world.
I can imagine someday just grabbing a 'pad' someplace, and telling it to give me the grocery shopping list, which will just pop-up at a little printer by the doorway as I leave. The items will all be cross-referenced against the fridge contents and cabinets, and me and my wife will both have our 'favorites' represented. Our food budget will have been consulted, as well as our schedules. Mostly the heavy bulk stuff will already be delivered by the weekly run, but because I'm more traditional, I'll still prefer to visit the store and pick out the fresh fruit and veggies. This week "fresh carrots and broccoli" from the little farmer's market, down the street. Nice.
No comments:
Post a Comment
Thanks for the Feedback!