> This, or something like it, is the future: the computing device is portable, and adapts itself to the forms of input available

Maybe it is, but it's not the future I'd like best.

I would like to have a desktop in the cloud, and data in the cloud, and nothing in my pocket; nothing to lose or break or have to make sure it still has juice.

When I get somewhere to work (a client's, a café, a friend's, etc.), I log in to my cloud desktop using whatever dumb machine is available, and off I go. Everything I do is automatically saved, backed up, versioned, synchronized for me.

A portable computer is useful in the subway, but at the office? What benefits do we stand to gain from still having things run or be stored on a local machine?

If you think your data "in the cloud" (which is just a way of saying "on other people's physical systems") is safe from theft, breakage or some foreign government agency pulling the plug, you haven't been paying much attention to reality lately.

And by "lately" I mean the past decade.

Don't get me wrong, I like the convenience of "the cloud". It's especially helpful in synchronizing all the data that I have on cheap and abundant storage on various devices.

We can have both. Why advocate giving up the one option that makes you independent from third party interference (or simply a lack of connectivity)?

We are talking about the future. In the future your data might be more safe in the cloud than on your own device, just like your money is more safe in a bank than in your wallet. It's possible.

Backup device manufaturers are really missing out on big bucks by not making a small NAS that downloads all you cloud data. Give it access to Google/Dropbox/Flickr etc an have it silently back everything up.

I hope they've thought of this and I've just missed it. They won't be able to sell USB based backup drives for much longer.

IOMega used to make a NAS with support for foldershare (the dropbox predecessor that was sold to Microsoft). That was 6 years ago - before it's time I'd guess.

The device 'Time Machine' by Apple does exactly that.

No it doesn't, it backsup your Macs over wifi, and that's all it does. It won't access web services, which is they key point I was trying to make. The Time Capsule, and all home backup devices I can name, make the soon-to-be-outdated assumption that your data exists on a hard drive you own.

Well more elaborately your mac can access all web services and download all data which can then be wirelessly backed up by the time machine.

For example my Dropbox folder is automatically synced on my mac which is then automatically backed up by the time machine.

Wouldn’t login sniffing on these supposedly dumb machines be a greater concern? Murders are more common within families after all.

While the I/O from those plugged-in devices could still be sniffed, including what appears on a screen, at least you could encrypt your communication when using a portable device.

> I log in to my cloud desktop using whatever dumb machine is available, and off I go.

Unfortunately there is no way to verify that that machine is actually "dumb". Fundamentally this is because modern computation and storage are microscopic phenomena. Even today we take it as an act of faith that our freshly produced electronics only do what the manufacturers say - and this is an ongoing problem for US defense contractors that outsource component manufacture to China.

This problem is also experienced by travelers who expect a "dumb browser" at the internet cafe, when in fact they are being key logged.

> What benefits do we stand to gain from still having things run or be stored on a local machine?

There are a few, such as statefulness that is maintained even without network connectivity. But the big one is that, presumably, "your" device is more trustworthy than a dumb terminal because, at the very least, you haven't acted to subvert your own security (although even this is often not the case).

If consumer networks were as reliable as roads and the electricity grid, why would I bother giving my data to someone so they can put it in a datacenter?.

The future I would like best is a device on the end of my home internet connection.

I could authenticate myself to other online services against it.

It could be a node in a social network (much like diaspora) where I can control all my own data.

It could have access to a NAS at home, allowing me to stream my favourite music and movies to wherever I happen to be.

In the future, with resilient enough networks, the datacenter will be distributed.

You described exactly what the people at http://www.freedomboxfoundation.org/ are trying to do.

It's still in an early state of development but you may join the mailinglist or read the wiki ( http://wiki.debian.org/FreedomBox ) if you are interested in this topic.

This is a great hardware startup. Sell rack boxes that can be installed by an electrician next to the power in/telephone in and get rid of those under featured consumer routers.

Hmmmmmm.... I'll sell you one for £500 a year, replaced regularly and managed?

I was thinking more of a Raspberry Pi, though obviously combining this device with the router does make sense.

Seriously? Why not combine into the NAS?

I could see a NAS needing to be upgraded more often than this device.

If the software and feature set were stable and mature enough then I can't see you needing to upgrade it very often.

Why not carry the dumb terminal in your pocket and drop it into a monitor and charging station at a client's, a cafe, a friend's, etc? That way you have local access to things that should be local (I don't always have Internet access on my phone here in Michigan, but I still want to listen to music), but cloud access as well? Phones are pretty well tied to the Internet these days anyway.

These two futures don't have to be mutually exclusive, and I'd argue we'll see them both on the same device at the same time.

Control, security, trust. Even if I got rid of all other computers, I'd still have a phone.

More and more I want to fully control my data on my terms. Maybe I'd be fine paying a service (linode?) to host my fully encrypted data in the cloud, but Google? Facebook? No thanks.

Whats to say the data isnt in the cloud? I expect that the future of computing has a few layers.

1. Device (basically CPU) - this could be your phone or something similar

2. Storage local cache on your "Device" but sync'd with the cloud in case of it getting lost or damaged

3. Input. Lets face it, we all enjoy a good keyboard and mouse...if you're on the go touch screen does the job but I'll never write a document from one (unless I am super desperate)

> What benefits do we stand to gain from still having things run or be stored on a local machine?

Security.

Do you store all your money at home or do you have some in bank? There's no reason why some big entity couldn't guarantee that the data you put in will get back to you whatever happens to company holding your data. If you mean security as in privacy, then there are cryptographically safe methods on preventing that.

I think the crucial difference here is the number of entities I have to trust. People can generally do a bit of research on a bank before they join one, if a bank really sucks you can switch, and if a bank seriously defrauds you then you know whodunit and can pursue legal recourse. By contrast if I'm using a supposedly dumb terminal provided by some random cafe then there's absolutely no way to prevent them from seeing everything I do [1], or from inserting arbitrary commands into any service I use. And if I use the provided computers at many locations, as you suggest, there's probably no way for me to figure out who it was that screwed me.

[1] You say "cryptography". I say "Analog hole".

I'm not talking about tomorrow but rather distant future. The banks evolved somehow. Now it is really hard to "become" a bank. There are very strict regulations in place. If a bank goes bankrupt the personal savings are still guaranteed to be paid off. So this is just a matter of making some regulations and establishing proper hierarchy.

And about security as in privacy. I don't know what future holds. Perhaps you'd be implanted a token that you can use for accessing the data which can't be faked or copied. Who knows. I just wouldn't rule out the possibility of a "cloud world" completely.

Even with strong crypto, if you're typing in your passphrase on a "dumb" machine that someone else owns and controls, you're risking the entirety of your security on how trustworthy (and security-competent) the owner of that machine is. No thanks.

You could have a token that would work exactly as credit card. Only one token would be authorized. You cannot copy it so if you lose it, you know you lost it can be "disabled". You would then need a new one with the same key however. That key could be stored in real bank's deposit box. Hardware of the token would only allow to store the key and not read it. I mean, that's a possibility for the paranoid ones who are not happy with cell phone auth for example.

I know there's no infrastructure in place (yet) but we're talking about some kind of future, not about today. It's certainly possible though. Even secured.

two-factor auth helps?

If somebody steals your money from a bank, the bank can replace it. If somebody steals your data from the cloud, there's no way to make things ok.

It doesn't matter that someone have a copy of your properly encrypted data.

>I would like to have a desktop in the cloud, and data in the cloud

You would sacrifice your freedom and privacy for a (dubious) convenience?

>What benefits do we stand to gain from still having things run or be stored on a local machine?

So that you won't be in paralysis when internet is unavailable?

Almost every answer emphasizes security, but what about convenience? Where is your dumb terminal when you climb in the Alps and wish too take a short at Mount Blank? In a taxi will you have to log in a machine before checking your friend'a address? Where is your dumb terminal in the washroom (probably the place where smartphones are most used)?

In those cases you don't have a big display or a dock either; you just use your phone and that's fine. What I'm questioning is the value of a dock+display to access your phone versus a "cloud" device that can be accessed from anywhere (including your phone).

I'm experimenting with that, actually: https://github.com/pfraze/link

Exactly. It would be great to be able to use an Optimus Prime or something as a mobile development machine. My dream for android development is definitely to be able to build and deploy on a single device.

Of course there's a way to go for that. The range of devices that support mobile docks is very small; I would prefer support for tablets with attached keyboards/mice. If they managed to get this working on tablets, I would definitely consider something like the transformer prime to replace my netbook.

I think it's better as a thin client. Already I can (and have) used ConnectBot on Android to connect to my shell and fix webserver (and content) problems from my phone, in the field (3G).

A full Ubuntu installation is a superset of a thin client.

Or in other words, a thicker client.

Androbuntoid: a THICKER client

I don't see why this would be the future.

I'm going to want to carry a "phone" for sure. Most people who've ever had a smartphone expect to have one the rest of their life.

I'm also going to want (at least) a machine with a keyboard and bigger display than my phone.

Why in the future would I want to tie one to the other? I wouldn't want that today — my computer works just fine without having to do some weird sync/plug-in dance with my phone when I want to use it. Even if this was seamless and even if we ignore concerns like finite batteries, what would I have gained besides a single point of failure?

I of course would like pervasive data sync. But why would I want the hardware of my computer to rely on the hardware of my phone?

I'm not sure you or I are the people most in need of a solution like this.

Consider, there are over 5 billion active cell phone subscriptions in the world. That's more cell phones than toothbrushes or toilets. And smart phones are being adopted in the developing world at a surprising rate. For many people, a smartphone will be their first and only computer.

If Canonical can enable desktop style content creation for people in emerging markets who will "leapfrog" over computers, just as they leapt over landlines, the impact could be enormous.

Why would someone in the developing world want their computer to be reliant on their phone? Presumably for cost, but that isn't an argument for docking-phones, is it? You'd still need the system case/power supply, display, keyboard, and pointing device (and would presumably also have speakers and camera, though I suppose you could reuse those of the phone, if you were really trying to cut costs). The only thing missing is the "computer", and an ARM system-on-a-chip that packages multicore CPU, memory, and a little storage can't cost more than $10 today, can it? Besides, since you'd have to have smarts to drive the networking/remote desktop features of the non-computer computer, you'd probably use a very similar chip anyway.

Why? To have all your computing stuff with you all the time. That is not just your documents, but your programs and settings. And to only maintain one system. They way i see it, the only rival to this vision is the cloud, and there are a lot of issues there[1]

Battery isn't going to be a problem either. With today's solutions you could just plug in one cable which connects to the monitor and charges the phone. Tomorrow- your desktop will have a surface to put your phone on that will charge it (like your toothbrush charges), perhaps authenticate with NFC, and then connect to your display with WiDi or similar, and your keyboard.. yes, seamlessly. Another thing you will have gained is you won't have to buy a second device. Of course, we here are likely to be the last to give up a computer-proper.

[1]: other than the obvious trust, reliability and privacy issues-- there is currently no solution for most applications-- licencing issues? many companies are likely to drag their heels here. And when there finally is a technically perfect and complete solution; how are you going to pay for this? advertising, $, or selling your data. Or, you could just keep it in your pocket, yourself.

Can you clarify the phrasing of your cloud issues? Do you mean there is no solution for trust/privacy issues, or do you mean there is no solution for licensing?

As for paying, this is a non-issue, you are going to pay either way.

There is no one i trust with my data more than me, and large projects have a poor history with trust/privacy.

When I talk about licencing i'm talking about the future: where you could use the cloud to access full fat programs like photoshop (after getting over any technical issues with remote installs). AFAIK there aren't provisions in current licences for such things.

Even if you have to spend for (more) local storage/backup, it is very likely to be much cheaper, and i would have a strong preference for the buy-once (and own) rather than rent monthly model.

So you are willing to commit to cloud-only, controlled by a random corp and subject to network availability but you are not going to trust the hardware you can carry in your pocket?

Notice that no one is saying that a secondary cloud-tier is excluded from this vision...

If I didn't want to do cloud sync, I'd be more excited about a local sync. One doesn't tend to generate gigabytes of data on a phone so even a Bluetooth sync would be sufficient for most people.

You don't use your smartphone to capture HD video, do you?

I do from time to time. Even they can be transferred over today's WiFi in a matter of seconds. My point is that keeping local devices in sync for all but a very few uses is already today easily solvable. So I wonder why we'd want to move to an architecture where our computers are reliant on our phones... there seem to be some significant tradeoffs to that model and it's not clear to me what real gains there would be.

>keeping local devices in sync for all but a very few uses is already today easily solvable.

No, it really isn't. Manual file transfer is... doable, but still more problematic than it should be. But there is absolutely nothing that will seamlessly keep my phone in sync with my tablet when I don't have network access (i.e. I take a photo on my phone and it's on my tablet, without me doing anything). If you have a solution, link please.

It's not solved, it's easily solvable. :)

Desktop PCs will always outperform portable devices. I think future docking stations will include both a big display and a big processor. When you dock your phone, your phone's apps will be running on the dock's fast processor while accessing your personal data off the phone. When you undock your phone, your running apps will seamlessly migrate to the phone's processor(s).

This will be extremely asymmetric multiprocessing.

"Desktop PCs will always outperform portable devices"

That doesn't mean portable devices won't replace them. Laptops began to replace desktops a few years ago, the same way PCs replaced mainframes.

That's right: once a certain threshhold of compute performance is reached, the excess is only useful for specialists and gamers.

The other element is that mobile devices are optimised for power consumption / battery life and also for heat output. Desktop devices generally have neither of these concerns, unless they strive to be 'green' or fanless products.

In the era of dark silicon, it might make sense to have those extra processors in the phone itself but only turn them on when you have an external source of power and a good physical connection for dissipating excess heat.

That's too complicated a solution. Linux has had excellent boot time detection of hardware for years now- dock your phone in the desktop chassis, hit reboot, the OS gets loaded from the phone's storage and runs on the tower's CPU and GPU.

Why even reboot? Treat the dock as a hot-pluggable processor: dock your device and the OS can migrate processes to the dock's faster processor.

Mostly because it's simpler- hot pluggable processors offer more and more varied ways of shotgunning your feet.

please leave the novelty accounts elsewhere

You might be right that desktops will always outperform portable devices (even though the gap is certainly narrowing), but mobile phones used this way are so convenient that the tradeoff is worth it.

The way things are going desktops are getting TOO powerful.

That's kinda the same as saying Mini-Computers will always outperform Desktop PC's.

The general trend of hardware is that it gets smaller, faster, and more efficient simultaneously. It is highly likely within 10 years you'll be walking around with a smartphone more than 10 times as powerful than the fastest 16-core desktop PC you can buy today, and possibly desktop PC's won't even be sold by then because there won't be any need for them.

If that's a bet, I'll take it. You'll still be able to buy desktop computers in 10 years. They might not be common, cheap or easy to find, but you'll be able to buy them. IBM still makes lots of money selling mainframes, and HP still makes and sells OpenVMS minicomputers.

Workstations at offices, gaming machines, animation and image manipulation will all probably stay on desktops and that will keep them around for years.

The scale of a desktop fits and works. Like how we could make a TV the size of a watch, but we don't because the large 60 inch plazma format, while immobile, has benefits worth the tradeoff.

Full featured mouse / keyboard, larger screens, surround sound are all perks of stationary desktops (I guess you could attach all these peripherals to a laptop, but then you are partially defeating the point of being mobile.

You really think desktop computers in an ATX case, with a full ATX-motherboard, and 5" hard drives are going to be around in 10 years? Seriously? They're becoming rare even today.

After reading that I looked down at the desktop I'm writing this from.

I laughed.

Desktops are by no means "rare". Cell phones have just become ubiquitous, theres a difference. (Laptops too for that matter.)

One of my problems with current generation laptops and netbooks (Besides costing almost twice as much for the same performance.) is that despite being "mobile" computers I end up having to set it down on a flat surface every time I want to use it. Walking around and using your laptop at the same time? Not going to happen. (Unless you buy one of those chest-strap things I saw in the first splinter cell...) The OLPC solved this pretty elegantly with the handle on the back, which is made just perfectly so that you can actually hold the computer with one hand and operate it with the other. (This mainly works because the OLPC is more netbook than laptop.)

And then you get two more problems I have with the laptop form factor: Trackpad vs Mouse, "Keyboard" vs Keyboard.

Trackpads, even when you have a responsive one, are awful. They're like touch screens, but not as cool. If you have anything on your hands, it gets on the trackpad when you use it. Trackpads are too slow if they're set at a speed where your mouse can actually point, and uncontrollable if the speed is turned up.

The "Keyboard"s that comes with most laptops barely qualify for the word. Traditional desktop keyboards can have keys with a height greater than five millimeters. (I'm exaggerating obviously, but laptop keyboards are very flat.) and I much prefer this to keyboards that feel like I'm tapping on a piece of plastic cardboard. I'd almost rather forgo the keyboard entirely and just push down on the little gelatinous circles underneath directly. And don't get me started on replacing one of those keys if they break. (Which they often do.)

And no, touch screens aren't the answer. On laptop screens, touchscreen == dirty display. I already have a rant here about touch screens on mobile devices. On displays bigger than a laptop, you better not have been planning to sit more than an arm's length away from your monitor. Oh right, the future doesn't have big screens in it.

Desktops might not even be common in a decade, but they'll be there. If for nothing else because theres people like me who still like to sit down to a (awesome) computer setup.

You can use whatever keyboard you like, whatever input device you like, but there is no inherent reason why your computer has to be an ATX desktop form factor and anchored to the wall. If you could accomplish everything you currently do on your computer right now, with your phone and plugging your phone into a base station, then there would literally be no reason to have a desktop at all. We are only about 1-2 years away from this reality.

Well. I'd say it's more like five years away from any sort of mainstream acceptance. Not that I'm some anti-hipster, but I don't think it will become mainstream right away BECAUSE:

1) People without smart phones exist. I'm one of them.

2) Nobody is going to be satisfied with phone hardware to control both their phone and their sit-down setup. At least not for a few more generations of Moore's law, and Moore's laws current method for increasing power, stacking cores, is a battery vampire. People who want their phones to actually run for a non-trivial length of time will end up under-utilizing their hardware while on the go. (Hardware that they still payed money for.)

3) So lets say we still have the sit-down setup in the form of the base station. This brings up it's own set of questions:

[A] Where is data stored? What if someone steals your phone? What if you don't have one?

[B] Where does the computing power come from, is this just a shell for the phone or does it have some power on board?

Now we can wrangle over these questions for a while, but I'm quite confident the answers are "The base has an internal HD" and "The base has computer parts inside." at which point I have to ask; how is this that different from what we have now?

I mean, sure my desktop is a little heavier to move than a "base" might be. But unless this base is going to be like the old 90's laptop bases (Huge flat thing with peripherals that plugs into a smaller top.) I can't really see myself moving it. If it has a full fledged keyboard, monitor, and mouse, I can't see myself moving it. If it doesn't, then it's a laptop. With all the problems laptops entail. (Which can only be fixed by "desktopifying" it.)

Of course, marketing conquers all. That is what everyones always going after right, true marketing? Or was that true love? I think it was the marketing...

If you're working with a keyboard and mouse, and looking at a big screen, and you have a rich multi-tasking environment that does what you need it to do, does it matter if it's running on a smartphone that's connected via bluetooth to a keyboard/mouse and via DLNA to a screen? Why the need to have a big box containing your CPU, if a little one will do?

Desktop PC's will always find a niche, but they're already a niche today. Laptops by far outsell desktops, and smartphones+tablets by far outsell laptops. Right now you have to be a geek to buy a desktop, because regular people don't buy those anymore. I don't know anyone "normal" who bought a desktop in the last two years. The idea to buy anything other than a laptop simply doesn't arise. This happened in 5 years. Why couldn't a smartphone-centric transition happen in 5 years as well?