[gjsman-1000]

Nope, it won't, as you say. How many people have Surface Pro X, or another Windows on ARM machine?

Less than 1% do. Let's say, though this is a high estimate, 0.5% of Windows users are ARM. Now let's say my app has 1 million installs at $1 each (pretty successful). That means my total value for optimizing for those ARM users is... $5,000. And they can already run my app with translation, so it's actually worth less than that. So... when you factor in costs of Project Volterra + Developer Time + Fixing Code Time + Lost Opportunities on More Productive Things... there's basically no way, even with 1 million users at $1 each, to justify the effort.

This device is great for developers who already care and love their users. Anyone who doesn't care won't care.

Edit: Also, at those numbers, it would be better if I didn't offer an ARM-native version. If even a tiny percentage of my userbase downloads the ARM-native version by accident on their x86 machine, the support costs will eat away at that too.

Edit @bartlettD: We're not - sorry if it reads that way. I'm just providing an additional argument, but I've reworded slightly to make that a bit clearer.

[jve]

It takes 1 CEO of Windows ARM device for dogfooding own software to ARM :)

Anyways, what I noticed is that CxOs started to carry Apple MacBooks around and sysadmins just had to deal with it.

[bartlettD]

I don't think we're disagreeing. Devs have no incentive to develop for ARM if there is no market share like you said, so they're not going to buy a $599 kit and its existence won't sway them either.

I think there will come a time when the number of ARM Windows machines increases though.

This kit is useful for the devs who care about the performance of their programs on ARM kit but they'll only optimise for it if the market is there.

[giaour]

> Less than 1% do. Let's say, though this is a high estimate, 0.5% of Windows users are ARM. Now let's say my app has 1 million installs at $1 each (pretty successful). That means my total value for optimizing for those ARM users is... $5,000. And they can already run my app with translation, so it's actually worth less than that.

Not disputing your numbers, but for devs writing web services or cloud workloads, compiling for arm can mean spending a lot less on VMs. If I were writing software that had to run on Windows Server on ARM, I would probably want one of these devices for local development.

[dboreham]

> How many people have Surface Pro X

<raises hand>