Cortanagate: Round and Round We Go

Microsoft has confirmed that only Cortana devices based on Windows 10 IoT are required to have a display. Meanwhile, those based on Linux or other systems can utilize a display too, but they are not required to do so.

This decision is in keeping with my philosophy that Microsoft just can’t stop punching itself in the face.

Understanding the truth of this matter required a lot of research, and in this case, I didn’t do a lick of it: Mary Jo Foley and Brad Sams both did the work of trying to get Microsoft to clarify what’s happening here, and I’m relying on their efforts here.

That said, because Microsoft has a curious issue with communicating clearly, and because we do not yet have live Cortana code to examine, there is no way—absolutely no way—to have a full understanding of the capabilities and limitations of Cortana on each supported platform today, despite Microsoft’s attempts at clarification. This is an all-too-typical situation in recent days.

(On a side note, yesterday’s Windows Hello episode is a great example: That tweet I referenced in no way mentioned Windows 10 Mobile, nor did any of the surrounding tweets. Point being, that wasn’t my fault, that was poor communication from Microsoft.)

So. Here’s the story in a nutshell.

Three days ago—and to be clear, it feels like a freaking eternity—Mary Jo reported that Microsoft was bringing Cortana to IoT devices “with screens” in 2017. This report was based on a WinHEC 2016 session which you can now watch for yourself on Channel 9.

I immediately dismissed this effort as both half-hearted—Microsoft should already have Cortana devices in-market now—and half-baked, as requiring a screen is a crazy limitation when you consider that the market leader, Amazon Echo, is successful specifically because there is no screen: It uses natural language interaction instead. Amazon has sold millions of ’em.

But then Microsoft made an announcement in which it teased a screen-less Harmon Kardon-built Cortana device. Microsoft, it seemed, was going to allow device makers to release headless (screen-less) Cortana devices too.

Brad asked about this and received the following quote from WE, Microsoft’s primary PR firm. It’s attributed to a Microsoft representative.

“The Cortana Devices SDK enables OEMs/ODMs to build smart devices powered by Cortana that are equipped with a microphone, speaker and/or displays. It’s delivered as a collection of open source libraries and protocols that can be used by OEMs / ODMs to embed Cortana on platforms of their choice including popular Linux variants and embedded OSes.”

This statement suggests the following:

Cortana devices will support a microphone, speaker and/or displays. And/or means you can have all of those or just some of those. And while a microphone and speaker (or at least audio-out) would seem like requirements, the “or” in front of “displays” suggests that that latter component is truly optional.

Cortana can be embedded “on platforms including popular Linux variants and embedded OSes. We already knew (from the WinHEC) session that Microsoft would support Cortana devices based on Windows 10 IoT Core, so this adds the notion of “Linux variants and embedded OSes.”

I will note that this quote does not in fact reference Windows 10 IoT Core at all. So does the support for “microphone, speaker and/or displays” only apply to “Linux variants and embedded OSes,” and not to Windows 10 IoT Core?

On yesterday’s episode of Windows Weekly, Mary Jo said yes. Using logic—always a mistake with Microsoft, I’ll never learn—I argued otherwise.

But as is so often the case—and, yes, this is a situation I’m growing comfortable with—Mary Jo was right. And she updated a second post on this topic today with a new quote from Microsoft, which reads:

“The Cortana Devices SDK enables OEMs/ODMs to build smart devices powered by Cortana that are equipped with a microphone, speaker and/or displays. It’s delivered as a collection of open source libraries and protocols that can be used by OEMs / ODMs to embed Cortana on platforms of their choice including popular Linux variants and embedded OSes.”

Yes. That is the same quote that Brad got. They’re circling the wagons over in Redmond, and Microsoft/WE refused to tell Mary Jo which operating system or firmware will be inside the Harman Cortana-powered speaker. Meaning, of course, it’s not Windows 10 IoT Core. Because, remember, Cortana on Windows 10 IoT Core will require a display. (Unless there’s a tiny screen we haven’t seen yet. You see the issues we face in trying to uncover the truth.)

Round and round we go.

To recap, and again, this is based on what Microsoft and its PR firm has said, and is not necessarily what we might call “the truth,” or at least the full story:

• Cortana devices based on Windows 10 IoT are required to have a display.
• Cortana devices based on other platforms can have a display, but they are not required to do so.

As to the “why” of this insanity, that’s unclear. Naturally, I have theories.

First, Cortana on Windows 10 IoT Core is clearly based on Cortana for Windows 10, and that implementation does require a screen because, for all intents and purposes, there is no such thing as interfacing with Cortana on a PC without using the screen. In fact, Microsoft is extending those screen-based capabilities as we speak.

Cortana on other platforms is new code and while it may seem more versatile, it’s more likely it offer a subset of what we can do on PCs, if only because they’re not assuming a screen. That is, device makers may have more component choices with Linux or whatever, but the resulting products might be less powerful, not “full Cortana.”

Maybe.