It assumes linear, progressive growth in line with what we see.
But technology doesn't work that way.

Namely, free stuff has always been, and will likely continue to be, a rising tide of stuff. Stuf that... well... you can get for free.
You can't sell DOS to a market where Linux is free, or Office 95 to a market that has free office products that cover most of the basic functionality.

The point I'm making is not "payware is doomed". Far from it. But it starts like a wild west of opportunity, but over time the rising tide of free stuff drowns out a lot of the noise, and it's only those that manage to keep their head above it and progressively innovate and get better that contribute to what ultimately becomes the "settled" market.

Mobile software is still in its wild-west hayday. But when things get popular (and profitable) the probability that some developer throws his guts into a free alternative rises. Let time do its thing. Let the pay-vs-free balance settle and the PC effect to take over.

Yes, iOS will always probably make more because Apple-ecosystem users have a more solid standing culture of paying for their software.

BUT beware anyone who picks up that initial growth trend, extrapolates it linearly and builds mountains of logic on that.
Because, if we've learned anything, our beloved tech industry doesn't like'em straight lines.

I'd be concerned about high-temperature cooking of electronics under back yard conditions.
Too many minute-quantity rare/hazardous materials go into their making.

Cook them, make them airborne, and you risk getting cancer for free.

I wrote a long post recently on this.

I think MS going this way will be a much bigger landscape changer.
Consider:
1. ARM is already powerful enough to run modern desktops (quad-core >1GHz parts). Nowhere nearly as zippy as core-based parts, but functionally sufficient.
2. Jan2007, the PHONE was a DEVICE. Dec2007, the PHONE was an APP. This is about to happen to many desktops.
3. MS announced metro will be underlying UI, but that a traditional desktop windows7-like interface can be launched as neccesary.
4.
i. Apple stack: SoC->iDevice[MOSTREVENUE]->OS->App Store,Developer Community,Cloud App Package,Global Carrier Relations

ii. Google stack (inc. Motorola): stack: Baseband->SoC->Android Device->OS->App Store,Developer Community,Cloud App Package[MOSTREVENUE],Global Carrier Relations

iii. MS (inc. Nokia) stack in 2 years: Baseband->SoC->Win8 Device->OS->Cloud App Package,Global Carrier Relations,DESKTOP developer community brought over by unifying mobile/desktop platforms.

Microsoft has a one-up tho (on Google, not on Apple, Apple's doing their version of the same thing): desktop-in-an-app.

We will, of course, judge by execution (microsoft doesn't have a particularly good track record of doing NEW stuff), but it definitely hauls their powerhouse strengths - desktop & Office - straight in the middle where the battle is still raging. Might even make them relevant again.

If you can do it simply with air can you do it with liquids?

Some points here -
1. A VERY substantial part of the productivity apps are made by microsoft. If they can compile Windows for ARM, they sure as hell can do so with Office, Project, Visio and what not. And once they move, well, vendors will too. Remember what happened when they turned UAC on by default on home PCs? Initially some chaos ensued, the vendors that were doing stuff outside your filesystem userspace miserably broke, and MS had to herd them all to step back in line, because Hey, grandma's computer now had permissions and sudo. They've managed this kind of thing in the past and can do so in the future.

1.5 When there's a windows for ARM, Microsoft's Visual Studio IDE has windows-arm compilation target and your clients want ARM variants of your software, why wouldn't you give it to them? (yes, more QA and another branch to maintain. Big whoop.)

2. I agree in that there won't be one UI to suit both touch and KB/mouse at the same time. It doesn't make sense.

3. The UI is not as thick a layer of the overall software as you imagine. You can have an ARM rig running an OS and/or a piece of software that has two UI's - one for touch, the other for KB/Mouse. KDE 4 has been doing this for a while. Your device will run the program, give you a (more cumbersome) UI if you're on a train and looking at your Win8 phone, but fire up into a UI that resembles Win7 once you slap it in its dock on your desk in front of the big screen and lay your hands on the keyboard.
Like KDE, I think an OS will be able to switch to and deliver the appropriate UI on whatever device it runs and making it capable of switching depending on the hardware in play at any specific moment is not, as a colleague of mine likes to say, rocket surgery.

4. You can also get creative. Like have the OS vendor's IDE compile every program or library twice to both targets, lump them together in a single universal binary and ship that to be everywhere. And have the OS break it down at runtime and only load to RAM the appropriate part. Diskspace taken up by your binaries is the least of your concerns. Then you can be running on either ARM or x86, have hybrid rigs that carry one of each CPU, hybrid CPUs/Soc's that can be either big-rig x86 or ARM depending on the environment and all manner of other cool stuff.

The slave and master analogy is a bit misleading because it lumps document storage and processing together and creates this false dichotomy of dumb powerless nodes vs thick clients. Reality just doesn't look like that anymore.

The phone will have processing (needed for an acceptable snappy UI, and because we need some form of CPU to drive it and even the cheapest/smallest ones are plenty powerful and growing by the day), will have local storage for OS and cloud-cached local storage (needed because the device cannot be assured to always have connectivity) and have authoritative copies of their data backed up to a cloud backend (beacuse, as you rightly pointed out, people lose phones).

The trick is to offer the cloud not as a per-app thing (where my contact list app accesses and locally cache my google address book) but integrated below the filesystem access calls (tho perhaps above the filesystem itself, or perhaps not)

If you've watched the June Apple keynote, Apple is driving in exactly that direction - extending the local filesystem and file storage API's given to applications to extend to their cloud. Expect Google and MS to follow, both know about OS kernels, datacenters and pipes and can do this easily enough, and they'll need to feature-match it.

X is just a conduit to stream a console over a network. Doesn't actually which apps you intend to stream and where are they going to come from.
Google docs is a decade behind.

The big question here is "can you meet, or at least approach, the 'WinXP, MS Office and misc X86 apps for windows' bar on Android running on ARM?"
It's the hardware to drive it, the OS to drive it, the applications and in many cases a way to tie into legacy stuff.

So here's the thing. Big tech is all about the verticals nowadays. Here's my future.

Apple showed us how it's done - having the CPU, the iDevice, the OS, developing carrier relations, an app store, a lot of apps and a developer community, and now a cross-device cloud service. Apple makes most of its money from the devices by the way.

Google's not letting down. After Eric Shmidt and Larry Page had their disagreement on whether Google should be fleshing out its own stack or consolidating around its "core business" (see Yahoo for why I believe that's a BAD strategy), Eric left and Larry went to work. They thought about their stack - same stack as Apple only top-down and with only part of the components - the cloud services, the OS, the app store and developer community, and its minor foray into the device business.

So they bought all the stack components they were missing in one lean and mean acquisition of virtually all 'things' Motorola - the solid carrier relations worldwide, a device making capability, the "defensive" patent portfolio - and they even one-upped Apple - they got another rung down - they now make the baseband too.

And here's where the big surprise rolled in.

Microsoft Windows 8.

Windows? In the mobile space? Weren't they late to the party? Aren't they dragging their feet with some distant relative of PocketPC? Wasn't their buddy Nokia about to be decimated and dismembered with cheaper ~350$ iPhone models and cheap Androids in some 100 countries with no carrier subsidies? You know, those places where Nokia still sells more phones than everyone else in the world combined? Those places where nobody buys 500$ phones?
Apple and Google are still going to take them to pieces, right? You know, Apple driving a cloud software package and "Cord-free" in those same countries where many people don't have enough money for both a PC and a phone?

Well... just hang on for a second and let's think about it like rational geeks who pertain to understand why Android and iPhone changed the market.
So don't sell your microsoft stock just yet. Looks like they've been thinking it through. REAL hard.

Remember how in 2007 a "phone" was a device? As was the music player, the GPS and to some extent the camera? Today, just like the others, the phone is an app. Sure, we call the device a "phone", but that's just legacy that stuck. Almost like calling a computer a typewriter. It's a rabidly multi-dimensional device. It's a web node, a tricorder, a content delivery platform and a bank terminal. And so much more. We're not in Kansas anymore, Toto.

All this to say, a phone is an app.

So it's 2007 all over again, only now it's Microsoft actually doing something /different/ for the first time in 24 years. It's their defining iPhone moment.

It's all in the PC, stupid - it got commoditized, all but forgotten, but it still does al the heavy hauling of our actual work.

  And on the new breed of mobile devices - "phones", tablets and whatever follows, it is, if Microsoft have their way, going to become, plain and simple, an app.
  And not just any app mind you. It'll be the killer app that will allow a lot of people to drop their desktop or laptop.

You'll hook your phone or tablet up to a screen and a keyboard (with or without cables), or not, launch said app, and do your word, excel, visio and other work stuff. When you close this app, under it all will be a mobile OS UI on-par (one would hope) with iOS and Android.

Cute, but where's microsoft going to conjure the entire stack needed to pull this off? It ain't a single-layer market anymore where you can get by as a big player making just the OS or just the device....

Microsoft isn't as bad as you'd think in their stack. They have handsets, basebands and carrier relations covered by their now best buddy Nokia, both them and Nokia have access to CPUs, they have a mobile OS that unlike Blackberry, webOS etc is actually competent, with a new kernel and the metro UI everyone's raving about, and they have the momentum and can wear the hard slog of gearing up an apps community and app store. And they're already pushing the cloud package with Office 365. Oh yeah, and they make their money from selling Office, so don't expect them to wither away and die anytime soon.

Apple can do this in a blink by the way. They have a powerful desktop OS they can just integrate straight into their mobile stack. They're already laying out the groundwork in fact - notice how you can show your iPad screen on an external display wirelessly? Notice how the "PC" was demoted below the cloud? Or how iDevices no longer require a PC tether?

Think how useful it'll be when your iPhone is running real desktop stuff in an app. And driving an external 30'' display and keyboard wirelessly.

And here's the irony - Google, seems to be missing a layer of the stack everyone, particularly themselves - has pretty much written off as "there's no business to be made there" - the desktop OS. And Chrome OS is not a desktop OS. It's a browser. Android can do what it does.

But the desktop hardware is going to go the way of the dodo, and the desktop OS will become the killer app of the mobile. And google will either need to somehow emulate for ARM-OSX (ya-ha, because Apple will so agree to that) or ARM-Win8, or see their hard-earned chunk of the smartphone market erode to Apple, Microsoft and Nokia.

And before you holler "There's not enough CPU or MEMORY BANDWIDTH in mobile devices!"... This is never going to happen at scale!"

Well, there's seemingly three CPUs out there, but there's really only two.

There's ARM CPUs with cortex cores - everything from the snapdragon, the reference A9 or Apple's A5 or countless others. That's the pick for mobile phones and tablets. Why? Because they have most of what we traditionally call a "motherboard" on the chip itself - a System on Chip. And the whole thing chews up 2 Watts. They're not particularly powerful compared to your desktop core i7, having nowhere near as much cache or predictive execution logic, but they'll take on an Atom.

An Atom also takes 2 Watts. Only unlike the ARM CPU, it doesn't have the entire system on the chip, and needs another 15 Watts of infrastructure around it on the motherboard. This makes a whole Atom system a relative power hog and takes up motherboard space not exactly abundant in small mobile devices. It's only competitive advantage is that it's a good old x86 architecture.

Which is why it's neither here nor there. And the moment ARM versions of Windows will appear, the Atom is dead weight. It can't do anything you can't do far better with an ARM. On the performance end, the CORE (and AMD) will chew it out. I'm writing these words on a "computer" that is as thick as an iPad, marginally longer, has a motherboard that can fit in a phone and drives a Core i7 CPU on a meager 17 Watts (TDP, less most of the time) for 5 hours. When a "desktop PC" killer app is out there, expect it appearing in a tablet (and its even lower voltage siblings in a phone) near you. And it will drive a working PC app juuuuust fine.

I'm not suggesting you can stick an i7 in a phone just yet. An i7 is not a system on chip either, and the system will at its core require no less than 30 Watts (and nobody will buy a phone that lasts 45 minutes, or want to hold something that emits as much heat as a 30 Watt light bulb). But with a killer app to drive demand, several years of development, using the equivalent of an i3 rather than an i7 and perhaps CPU designs that haul the memory controller onto the chip, alongside higher capacity batteries... it's not science fiction either.

My point in all this discourse is that while the OS's will take a few years to get there, and as the killer app generates the mass demand, ARM, core and other architectures will clamp down on the gap in the middle, expect something powerful enough to drive desktop, and power-cheap enough to get shoehorned into phone form factor to come on offer. Some will remind you we were driving desktops with XP and office >2000 on machines with 450MHz CPU's as little as a decade ago.

And why Google just wagered on the Atom in its new Intel pact is beyond me. Seems like the wrongest route to take given it is already ill positioned to cope with mobile devices that will drive desktop OS's.

I think it's still their move - everyone else has made their long-term play that seems to be making sense, and they need to show how they intend to swim in what is now a virtually certain future, at least insofar as the other two major players have placed their bets and are plowing their way relentlessly in that direction.
  It may still be a good 2-4 years off, but it's now officially been put into play.

Interesting Tmes.

The average gamer, as identified through PAID software demographics, is 37.

Put otherwise, the average PAYING gamer is 37.

That's some good ideas. Hope Gabe gets around to reading this (I've emailed him, both with the original post quoted verbatim, and also linked this sub-thread and recommended he go through the comments as well).

One concern I had after reading your suggestions is 'scope of penalty in case of abuse'.

Let's say you manage one billing account, and three gamer accounts - you and your two kids.
One of your kids does something seriously unsocial in an MMO - say he gets caught cheating or something similar.

The system should affect his gamer account adversely, but should NOT affect the billing account and the family library in a way that would impact the other gamers without giving some form of "slap-on-the-wrist" warning first (say, a three-strikes-and-you're-out system). Voiding/Cooloff on the kid's gaming account is useful for offenders who care about their account etc, but not so useful to the cheater that would just create another account under his billing one and go at it again. You do need to ultimately be able to penalize the billing account, but you also need to show consideration that there are other not-guilty people there too and let the guy who manages the account some room to maneuver, specifically if you're pitching to parents. If you want them to trust, respect and pay you, you need to convince them you're on their side.

If the store voids the entire billing account or the game library on a single offense (serious as it may be), and my other kid - a 5yo - comes back crying because he just had stuff taken away for something he did not do, you probably won't (ever) see me (the guy who pays) coming back to do more business because I'll perceive you as irresponsible and inconsiderate, not only towards my innocent kid, but to me by making my life as a parent harder than it needs to be. Sure, I'll take care of the offending kid, but creating an unnecessary front for me with the innocent kid means I don't respect the store anymore.

Flagging the billing account with a strike, and locking the account out only if you get more than so-many strikes inside a, say, 3-year period, seems reasonable enough to keep the "serial offender" billing accounts at bay while giving the families enough slack to retain them as satisfied paying customers over the long haul.

A friend microbiologist calls this an "evolutionarily stable strategy". Something you can do for a long time without the occasional statistical blip blowing you out of the water in an irreparable way.

It's not hard. Just have a look at how most first-world countries regulate driver-license suspension on traffic offenses.

Thank you.
Sent.

>> How would steam prevent you handing that one family account to several households and letting 15 people play under it at once?

By limiting the number of steam accounts per billing account to a reasonable number (say, 5 or so, it's at their discretion), and having a fee (that would make it a same-benefit-with-more-hassle option for the public you named).

The whole model is a cross-industry standard. Just like, let's see:
* Managing multiple cars under a toll-road family account via a web interface?
* Managing multiple SIM cards under a cell company family account via a web interface?
* Managing multiple bank accounts under a single "person" account?
* Etc Etc Etc

Everyone everywhere has long since come to the realization that "in-family" micromanagement
[a] Is a requirement and needs to happen, or else consumers will go get their service from someone else who does offer it
[b] It's stupid to employ call center people and expensive/cumbersome mechanics to do it when you can just give people a web interface and reasonable freedom in managing their own corner of the world in a way that doesn't interfere with business.
[c] If you respect consumer's time and make their life easy, they respect and like you.
[d] A family account with multiple members, uninhibited by cumbersome mechanics, has less snags between wanting something and paying for it. It ultimately spends more.

Simple
If everyone is selling a tomato for 1$, and historically it's been sold for 1$ (tho in the past with 5$ packaging now made obsolete by technology), making people pay 1$ for it multiple times based on an flakily-enforceable pseudo-honesty system is unrealistic and ultimately bad for your business.

1. Many people will opt not to pay you multiple times (in their mind, they paid for it once, they rightly deserve to be able to do with it what they can do with the 1$ offering from every other competitor).
2. If you somehow manage to semi-identify these users and allow them to still use the tomato but get a somewhat degraded user experience... that degraded user experience just became your de-facto product to every frugal parent who has more important things to do with his money.

It's just bad business.

Shai Agassi of Better Place fame made a good related point - that Coca-Cola is a 99% margin business.
Some very bright individual in history decided to use bad sugar rather than good sugar in their secret recipe, saving a bit and making it a 99.3% margin business.
That decision cost Coca-Cola mountains of popularity, with many people who could have been potential clients outright avoiding the drink altogether.

You need to know where you're saving and where your destructive bean-counting habit is costing you single or double digits in next year's growth.
Sometimes, it just makes more financial sense not to.

Something you may be aware of is the increase in popularity in gaming over the past several decades.
That translates into more households with more than one gamer, and more households with more than one gaming generation.
I game, my kids game, I have many friends whose partner games.

As an individual steam user, I find your prices generally reasonable, your service adds enough value (ignoring ethics and judging strictly from a convenience perspective) to justify paying you and using it over the hassles of both piracy and retail. Good job to you and your team on getting (me) there.

However, I, like many geeks of my generation, have now evolved into a family of five, and am no longer an individual steam user.

This is where the problems start, and you push me, your customer, away. Why? Because I'm a dad, and my gang all play.

For the sake of making a point, I will ignore 'offline mode' because the games we care about are online.

Here are the options you give me:

Option 1. Have one steam account per person, and either buy many copies of each title
(or, I am told, go through a cumbersome process that costs 10$ processing fee to have your support move the title between accounts, this option is too painful to be practical. ).

Insisting I have a separate per-game license for each kid makes sense and is fair if we will be playing concurrently (and it is A-OK for you to sell us a 'borderlands 4-pack'. I'll buy it.).
This makes no sense if I'm done playing a game, uninstall it, and my kid wants to have a go. Realistically, you're dreaming if you think you'll get me to pay twice. You'll either give me a way to let my kid use it, or I'll take my business elsewhere to GOG or direct2drive or retail, because they will.

Option 2. Have one account for what I'll tell you is /me/, but what in reality will be the whole family. I won't tell, you won't know. Sadly, that means that two computers on my home network can't be "on steam" at the same time, and I can't play online game X while my kid plays online game Y. Plus, it'll get all my steam achievements gunked up with my kid's ones. I don't want that. Force me down this route and, again, I'll go.

Option 3. I'll create a separate steam account for every game I purchase. This will make your product into a very inconvenient one with a flaky user experience, no achievement history etc, and I'll take my business elsewhere. Too much hassle.

Here's the news. An entire gaming generation is now very busy having their children reach gaming age.

You can put some weight behind those brave words you said. The solution is dead obvious.

The recipe is:
1. One family "billing account" (that's a BILLING account, not an application account you sign into steam with) with a single billing method. If a single billing method isn't enough to deter most of the unrelated people from pooling into a "pretend family" account and costing you potential revenue (it probably would be enough, and while you may lose a bit of immediate revenue, you will make huge gains in customer loyalty by trusting them), then put your thinking cap on and figure out how to structure a plan to include real families that count money together and exclude most of the freeloaders. You have smart people working for you.
2. ONE family-wide game/license library.
3. Several "gamer" steam accounts, one per real person managed by the billing contact (the guy with the credit card who vets the games, aka the parent), without needing to involve you. That's what web interfaces (or your application) are for. These steam accounts should all be able to go online concurrently, and can all have their own (SEPARATE) steam achievements, and can be use different games at the same time. If they want multiple people to be playing the same game at the same time (that thing we call co-op play is very popular in families btw) they need to purchase and own multiple licenses. Keep 2-pack, 3-pack and 4-pack deals coming.
Yes, this will mean you may have several steam achievements gained by a single game license. Deal with it. It's not like it costs you money to manufacture them.

4. One "parent/family-administrator" interface that can
[a] Vet games (parents might not want kids playing specific games in the library, violent etc)
[b] For those of us with more than one kid, decides which kid gets to use which title.

Now, this structure is obvious, and if you won't do it, someone else will.
And I represent an ever-growing segment of the gaming market who, unlike the kids, has a platinum credit card and money on it, and is not afraid to use it.

I challenge you to present my family, and every other family like it, with a good service we will like and can justify paying for.

The basic premises of the Matrix is fundamentally flawed.
Why the ^%$^ would you grow humans when you can grow, you know, YEAST, for much more benefit at a fraction of a hassle?