Hence Continuum.

But beyond that, UWP apps are far friendlier on smaller screens. Different screen sizes certainly demand different sizing and presentation of UI elements, but that does not mean you cannot serve different form factors with a common code base. Writing a ground up application for a handset is a much different proposition than doing a marginal amount of coding to make an existing one work on different screen sizes (especially since apps need to do that to work well in windowed or tiled mode anyway).

Once upon a time they got rid of two proprietary* standards (ADB and their mini-DIN8 RS-422 serial port) with a cross-platform standard. Now they're removing a standard port and offering two proprietary standards instead (Lightning and W1 wireless). Not quite parallel.

* Neither RS-422 signalling nor the mini-DIN8 connector are proprietary in and of themselves, but that particular configuration was, to my knowledge, exclusive to Apple. Sun also used mini-DIN8 on machines like the IPC, but the pinout was completely different.)

You seem awfully hung up on the initial characterization of Intel's announcement. They are not discontinuing existing SKUs. They are not discontinuing the Atom microarchitecture. They are not discontinuing the Celeron and Pentium lines that are already used in sub-$200 devices. They are not even discontinuing the Atom brand as it turns out; they just demoed the new line at the Developer conference this week: http://www.anandtech.com/show/10576/more-details-on-broxton-quad-core-ecc-up-to-18-eus-of-gen9

And while Microsoft may or may not decide there is enough market for a cheaper Apollo Lake based Surface, HP has already leaked details about an Apollo Lake based Pavilion x360.

Actually, I specifically talked about the "Surface line" and the "form factor", rather than the Surface 3 itself.

Likewise, the statement I was responding to did not reference the particular chip, but argued against the interface, the form factor and what you deemed to be unnecessary compatibility with the x86 ISA:

"I never saw the point in Surface as I found the Win8 interface unwieldy (and _unwelcome_ on my laptop), x86 compatibility unneeded in a tablet and the whole package expensive compared to tablets that did what I needed them to do (& not what Microsoft thought I should be doing with them).

I can reply for clarity, too. :)

That's why I used the word "effectively". For the lay individual, how the carrier lowers their throughput is an ineffable implementation detail; the important question is now how, but when and why.

It's not easy to compare the prices on products that haven't been released and have no public pricing yet. They certainly won't go as low as current generation Atoms, but I would be surprised if they don't offer options cheaper than current generation Celerons. Even if they don't, those have a "recommended customer price" of $107, which is well above the wholesale price to systems integrators. You can buy a Chromebook using one for under $200 without much shopping around.

It could hurt them long term in the ultra-budget category if they don't introduce something more modestly priced to replace Cherry Trail eventually, but there is not much margin nor growth there.

The majority of convertible devices don't use an Atom.

This is not actually going to be the only plan option. Even if they stop offering the Simply Choice plans (which they say they aren't) the pre-paid plans still offer service much cheaper. Individual plans start at $30/mo, and you can get a family plan for as low as $100/mo with 2GB per line, or $120/mo with 6GB per line.

They will lower QoS on lines that go over 26GB of data, which will effectively throttle their speeds when there is a lot of traffic.

They're not eliminating the current Simple Choice plans at this time (even for new subscribers, as far as I can tell), nor are they eliminating their prepay options under the T-Mobile and MetroPCS branding.

No one knows what the pricing Apollo Lake processors will be yet, but considering they will be using the same CPU and GPU core as the cancelled Willow Trail platform (the tablet variant of Broxton), it seems premature - and more than a bit silly - to conclude it will double the price of tablets.

Their mobile portfolio was looking redundant and unwieldy. Had they followed through on the roadmap as stated you would have had the Willow Trail SoCs (presumably in x3, x5 and x7 variants again), the Apollo Lake SoCs (with both 2 and 4 core options with both Celeron and Pentium brandings), the Kaby Lake-Y processors (also in 2 and 4 core configurations), and the Kaby Lake-U processors. Considering the similarities between the Willow Trail and Apollo Lake platforms, the consolidation is probably for the best.

They /did/ abandon the handset market, but given that the only logical market for that was for Windows phones, it's hard to see what their market for the part would have been anyway.

Considering the form factor is the only one showing significant growth, while both traditional tablet and PC sales flag, I suspect your assessment of "what most people want or need" is mostly projection.

You missed a couple key words in that "quote". Highlighted below:

"The resources dedicated to the two Atom chips will now be focused on products that can advance the firm's strategy and generate higher returns."

As I said, they are not dropping the microarchitecture. They are dropping the SoFIA and Broxton SOC platforms, but "will continue to support our tablet customers with SoFIA 3G/3GR, Bay Trail and Cherry Trail now, and later with Apollo Lake and some SKUs from our Core processor family."

Apollo Lake, of course, being the forthcoming Goldmont (i.e. Atom) chips released under the Celeron and Pentium branding. (This is nothing particularly new; the Baytrail and Braswell lines also used Atom cores and Celeron / Pentium branding.)

Regardless of what you call it, there's a difference between not seeing the point in a particular form factor and not preferring it for your own purposes. Are you really unable to understand why other people like the Surface line?