Now that Microsoft is in the tablet business with Surface (I don't say 'now Microsoft is in the hardware business' because Microsoft has been in the hardware business for three decades, with mice, keyboards and Xbox), the speculation that Microsoft will buy a phone manufacturer and bring out its own brand Windows Phone has started up - again.
But although I was very surprised that Microsoft is making its own tablet, I still don’t believe Surface makes a Microsoft phone any more likely, because the situation with the OEMs is very different in the tablet space and the phone space.
Put simply, Microsoft doesn't need to compete with its hardware partners to make a great phone because Nokia is already doing that - unlike the PC OEMs who have barely created one usable tablet PC design between them.
And Microsoft can't afford to compete with its hardware partners in the phone market, because they have too many other alternatives - unlike the PC OEMs, who have no serious alternative to Windows.
Certainly the combination of power-hungry processors, mouse-oriented interface and generic tablet designs has left even tablet PC fans like us disappointed.
It's that kind of sloppiness that I assume drove Microsoft to build its own premium tablet. And it's been a good thing for Windows as well; making its own tablet will give Microsoft a focus on making the hardware work as well as the software, on optimising the software experience for running on actual rather than theoretical hardware and on paying attention to every step of the process from setup to troubleshooting (because every problem on a Surface is going to turn into a costly support request for Microsoft directly).
None of that is true for Windows Phone. The limited hardware platform simplifies development and support even though there are multiple suppliers. Bundled apps can't slow the phones down like PC crapware can and you can get rid of operator and OEM apps you don't want. The hardware isn't a problem. Samsung's Windows Phones have been perfectly acceptable, even if they don't get as many updates as they should; HTC has a reasonable range of handsets covering a mix of price rnges and operators. And Nokia is making exactly the kind of innovative, well-designed, full-featured handsets Microsoft wants for Windows Phone, taking advantage of all the software features.
So Microsoft doesn't need to buy a handset maker, or have one make a Microsoft branded phone (the way Toshiba made Zunes) to get the devices it wants.
And it can't afford to alienate all the other handset makers it doesn't buy - because Microsoft needs the phone makers more than the phone makers need Microsoft.
If they don't want to make Windows Phones in competition with Microsoft - phone makers have plenty of alternatives. There's Android, there's Tizen (the descendent of Meego and several other mobile Linux projects), there's Mozilla's Boot To Gecko - you could even make a case for picking up WebOS and making a phone with it, given that the HTML5 underpinnings are more relevant than ever now.
The desktop operating system market is far narrower. If the PC makers are unhappy at Microsoft makings its own tablet, what are they going to do? Abandon the Windows marketplace where they sell millions of PCs and switch to Linux? No: even if you loathe Windows 8 with the unjustified fire of a thousand burning suns, this still isn't the year of desktop Linux.
Android tablets have hardly been a commercial success (unless you're Amazon) and with Google prepared to give away any margin it might be getting on the $199 Nexus tablet as a credit for apps from Google Play it's going to mop up most of the buyers looking for an Android tablet that isn't the Kindle Fire. Chromebooks still aren't a serious alternative either; they're far too limited (and still overpriced compared to say, a Windows notebook of the same spec running the Chrome browser).
Like it or lump it, the PC makers are going to keep making Windows PCs.
With no need to buy a handset maker and good reason not too, why would Microsoft saddle itself with a loss-making company like Nokia - which is already doing everything Microsoft needs it to - or RIM?
If Microsoft is really talking to RIM, I'm prepared to bet* it's about taking a licence for Windows Phone rather than buying a company Microsoft already has significant partnerships with in all the important areas (Bing and Exchange).
I'm not sure RIM is interested; BlackBerry 10 aims to take advantage of many of the same trends Windows Phone does and changing horses in midstream is a good way to get drowned. Yes, delaying BlackBerry 10 until next year is a huge blow for the company (Nokia is already putting the face replacement group shot feature RIM demonstrated in the spring onto current Lumia handsets) but they could hardly get a Windows Phone out before then anyway.
And while I'd love to see a Windows Phone with a BlackBerry-quality keyboard, the 4:3 format screen you'd expect to see used on a keyboard phone was noticeably missing from the lineup of Windows Phone 8 screen resolutions - and RIM is concentrating on touchscreen phones with clever software keyboards.
As for those stories about Microsoft looking at Nokia's books? Look at the significant cross-investment between the two companies. Microsoft is pinning its phone hopes substantially on Nokia and investing heavily in the company to make it happen. It's going to have been looking at Nokia's books in great detail to make sure Nokia has at least a chance of delivering. As always, following the money with Occam's razor in hand will help you work out what the smart move is.
* I already have a bet on twitter that Microsoft will not buy RIM, or Nokia, as a whole, functioning phone company: if either company got broken up for spare parts Microsoft would be in there for the patents, but if Microsoft buys a phone maker and makes its own phones, I’m buying the finest latte in Putney.