Ok, rumors out today that Microsoft is considering releasing Microsoft Office on Linux.  Cue immediate reaction that ranges from skepticism to outright hostility that anyone would repeat such a rumor.  And the most negative reactions are coming from people I respect.  Personally I think that both the assumptions about what an “Office on Linux” would be and all the hostility around the rumor are misplaced.

Now do recall that I considered porting SQL Server to *nix at a couple of points.  So I have first hand experience with taking this kind of idea to Microsoft’s senior leadership, including Steve Ballmer.  It is not the knee-jerk negative reaction that outsiders expect.  It is a rational encouragement to make the case.  Have Steve’s views changed in the many years since I talked to him about porting to a non-Windows OS?  No doubt.  At some points in the intervening years I’m sure he’s been less receptive to the discussion.  But in his efforts to remake Microsoft into a Devices and Services company I would venture he’s become more receptive than ever to such proposals.  Services need clients.  Services can not be allowed to fail because you refuse to support the clients that users actually use, even if they aren’t your clients.

Near the top of the list is transitioning the Information Worker business into the services world with Office 365.  The importance of Office 365 was clear when I was still at Microsoft, and the emergence last year of the Office 365 Home and Student Premium preview confirmed to me that if anything its importance continues to grow.  Don’t think Microsoft would make tradeoffs that weren’t in the best interest of its Windows’ business in order to accelerate Office 365 adoption?  After Steve Ballmer declared “we’re all in” on the cloud I (with senior executive approval of course) pulled committed functionality for Windows Server 2012/Windows 8 so I could shift the resources to things critical for accelerating Office 365 adoption.  This was after WS2012/W8  planning was complete.  In fact, this was after the first development milestone was complete.  And my team wasn’t the only team to make significant (and sometimes painful) shifts to support Office 365 (and Azure).  Steve wasn’t kidding around when he said “we’re all in”.

So when rumors of Microsoft considering bringing Office to Linux surface I don’t discount them so readily.  They’d be stupid to not consider it, which doesn’t mean it will happen.  It is important for Office 365 to support any and all popular clients.  That doesn’t mean each client has to be supported at the same level of depth or breadth.  Not every Office application has to come to every client.  The applications on each client don’t have to necessarily have the full functionality of the version available on Windows.  And the delivery mechanisms don’t have to be the same.

Office 365 supports Linux today via Office Web Apps, the Outlook Web App, POP3 (so you can use a local mail client), and (I believe) the Lync Web App (for meetings).  Is this sufficient?  I doubt it.  Take the simple scenario of an engineer (using Linux) collaborating with product management, marketing, finance, etc. on a business plan for a new product.  The later are likely all using full Office on Windows.  Can the engineer fully collaborate in document creation using the Word, Excel, and Powerpoint web apps?  Doubtful (unless everyone else reduces their use to things that work with the web apps).

Today’s solution for Linux users more often entails forcing them to use VDI to access a Windows Desktop with full Office, dual-booting into Windows, or running Windows in a VM on their Linux system.  The Linux users I talk to absolutely hate this and try hard to minimize how often they do it.  In a pure packaged product world, particularly with this small a user base, the strategy makes perfect sense for Microsoft and is acceptable to their customers’ senior executives.  But in a services world it doesn’t necessarily fly.

To begin with it turns the Linux users into advocates for corporate-wide adoption of Google Apps at the very time their CIOs are making a Google Apps vs. Office 365 decision.  In any enterprise the Linux user base may be tiny, but it has influence many times what its size would imply.  Generally the Linux user base is likely to include IT employees, placing them close to the decision makers.  In some enterprises the small Linux user base might include their most critical employees, such as the engineers in an aerospace company.    It would be silly to expect this user base to become advocates for Office 365, but that isn’t what is required.   Microsoft needs to be able to prove to the C-level executives that their Linux users are not excluded from an Office 365-based solution.  And they need to be able to do it to an extent that a rational impartial observer would agree with them.  Put another way, they need to be able to neutralize an argument from Google that it has a better solution for companies utilizing a wide array of platforms.

This is mostly the same argument for why Office for iPad (or Android tablets) is needed.  The difference with that argument is that most iPad users are already using Office on their Windows  (or Mac)desktops and notebooks.  And they want it on their companion devices as well.

When I hear “Office for Linux” what pops into my head is not yes/no, realistic/not, bad for Windows, etc. but rather “what exactly does that mean?”  Does it mean a port of the full Windows apps to Linux?  Does it mean as a packaged product, or only as part of the Office 365 service?  Or are they some new subset apps, such as what they might be working on for the iPad (aka, have they written a subset of Office that they can adapt to multiple platforms as Office 365 clients)?  How about greatly enhanced Office Web Apps that are only available as part of Office 365 (i.e., and not the free Skydrive offering)?  Other options?

Personally I think everyone translates “Office for x” into full ports of Office for platform x.  But I doubt that is in the cards.  These other platforms, be that the iPad or Linux, are likely to get subset offerings targeted at the Office 365 service and usage scenarios Microsoft prioritizes.  For iPad’s or Android those are companion device scenarios.  For Linux they are probably collaboration or general corporate citizenship scenarios.  On the upside this makes the rumors far more credible than most people give them credit for.

On the downside it means that user expectations are too high and the reality is bound to disappoint many.  Take Office 2013 for Windows RT as a simple test case.  The lack of Outlook has caused significant outrage.  Lack of support for old macros causes some old-timers to claim it isn’t real Office.  And sure enough the first comment I read about the Office for Linux rumor was from someone saying they weren’t interested unless it included VBA support, which of course it wouldn’t.

Still think it’s all about Windows and that’s why Microsoft will never bring Office to X?  Ok, let me accept your premise and propose why it is wrong.  The move of most enterprise app clients to a web model, the growth in acceptance of non-Windows devices in the corporate environment, and the BYOD trend have greatly weakened Windows’ hold on the enterprise client market.  While having Office support, any Office support, on non-Microsoft clients may result in some slip in Windows market share an enterprise shift from Microsoft Office to Google Apps puts the entire Windows client population within that enterprise in jeopardy.  In other words, it is in the best interest of the Windows business itself for Office to support other clients if that is what it takes to keep customers from moving to Google Apps.  Risk a 5% market share loss to avoid a 95% share loss? That is the real question for Windows.  (BTW, Microsoft overall the net impact would be positive as the increase in Office revenue and profit likely greatly exceeds any negative impact on Windows.)

So rather than sitting here and saying “it will never happen” or dissing those who are publishing the rumor I’m contemplating what the rumor might mean.  It could indeed be total BS.  But just as likely, maybe more likely, it tells us that Microsoft has something up its sleeve.  But it isn’t likely to be exactly what people think it is.