Predictions 2009: Microsoft Becomes Open Source Leader

January 5, 2009

My prediction for 2009? Microsoft announces that it is open-sourcing key components of its core office and operating systems.

Ludicrous, right? After all isn’t Microsoft the antithesis of open source: the Voldemort to Open Source’s Harry Potter?

Well no. A few salient points:

1.  Microsoft isn’t stupid. After all, it’s been one of the most successful companies ever, and it has historically done its best work when  under pressure: Windows 95 faced up to the Mac; Excel when faced with Lotus 1-2-3; Windows NT when faced with Taligent and OS/2 (once IBM took it over). Microsoft is currently faced with attacks on its OS server market (Linux, mainly Red Hat); its office suite market (OpenOffice). The desktop OS market is not far behind: the latest release of Ubuntu (Intrepid Ibex) is easier to install than Vista, is quicker and is less irritating to use.

2.  Microsoft has historically backed multiple horses, and quietly dropped the ones that weren’t doing too well. Remember Word for DOS? A much better word-processor than any version of WordPerfect, but phased out when Word for Windows started to succeed.[1] Multiplan? A DOS-based spreadsheet which was actually pretty good even when compared to Lotus 1-2-3, but dropped when Excel took over. It is easy to assume that MS dropped the DOS products to let Windows succeed, but if you look at Microsoft’s advertising during the relevant period (the late 80s to early 90s) it is clear that it was promoting the products as hard as each other. It was by no means clear that Windows would finally win, and Microsoft was keeping all its options open. With hindsight, a Windows win looks inevitable, but Microsoft was savvy enough to back both plans just in case. More recently? Despite having a ‘computer on every desk’ as its slogan for years, and therefore ostensibly rejecting thin client computing, Microsoft introduced a great web app in Outlook Web Access, bought Hotmail and integrated it into MSN, and licensed Citrix winframe technology to produce Windows NT  terminal server edition, all in blatant defiance of the ‘computer on every desk’ paradigm. And Microsoft is now promoting Office as a suite of web apps (responding to a Google threat). It even sold Unix, in the form of Xenix (which later became SCO/Unix in which Microsoft had a large equity stake), for years. So although it is easy to think with 20/20 hindsight of Microsoft’s plan being consistent and pre-ordained, in fact the business plan has always been to adopt a number of different models and competing products, and then run with the ones that work. If you’ve got the resources, this is clearly (and demonstrably) a pretty good business model. So for Microsoft to run open source and proprietary software in parallel would be entirely consistent with its previous business model.

3.  Microsoft is versatile – surprisingly so for such a large company. Remember when MSN was at its heart a communications network which ran in parallel with the Internet, and was incompatible with it? MSN soon realised its error, and ported MSN over to the Internet, and made Internet access an integral part of Windows 95.[2] Windows NT (much of whose code still underpins both current servers and desktop operating systems) was written without any consideration of Internet networking protocols,[3] and Internet e-mail compatibility in the early versions of Exchange Server was nothing short of a kludge: the internal workings were based on the X400 protocol, and converters translated between X400 and SMTP (the standard internet mail protocol).. In short: Microsoft has made plenty of decisions which could be regarded as mistakes, but when it has realised those mistakes, it has been pretty agile at addressing them.

4.  Vista is arguably the biggest proprietary software project that the world will ever see. How so?  Vista has been one of the largest software projects in human history, with 50,000,000 lines of code. It is debatable how much of a failure Vista has been (distribution stats seem to include licences which are then backgraded to XP and so are highly misleading). However, only a company with the deepest pockets could finance such development, and no sane management team (including Microsoft) would consider embarking on such a project again when open source provides a design and development methodology that is far less risky and inexpensive, and is proven to be just as effective. Bear in mind, also, that Microsoft possesses a customer-base which is astoundingly effectively locked-in, and even with all those egregious advantages, Vista has been (and will, I predict, continue to be) a disaster. There will never be another piece of purely proprietary software on the scale of Vista (Windows 7 will, I predict, contain a number of major open-source components).

5.  Microsoft has already dipped a toe into the world of open source. Its first foray, shared source, was pretty disastrous, and wasn’t open source at all: it was a means of getting to look at the source code, but not to use it, and Larry Rosen has cogently argued that this was in fact the antithesis of openness: by allowing people to see the code, it made it significantly easier for Microsoft to commence infringement claims against participants on the grounds that as well as the motive they now had the opportunity to infringe.

6.  Microsoft’s business model has to become more like open source companies: the HPs and Red Hats of this world. Absent some huge leap in usability or functionality (which is conspicuous by its absence), there are fewer and fewer improvements from each subsequent release of Microsoft software. Windows 2000 was a pretty good desktop OS, and it was difficult to persuade people to move to XP.

7.  Microsoft has used open source code extensively. Internet Explorer was based on NCSA Mosaic, and Microsoft has used (legally, I hasten to add) significant portions of open source code in its networking protocols, and elsewhere.[4]

8.  Microsoft started by ignoring free and open source software, then dismissing it, then attacking it as ‘viral’, and now, seriously embracing it: it has had licences approved by the OSI,[5] and has a number of relevant projects.[6] So to follow these Gandhi-esqe steps to their logical conclusion, open source has ‘won’ at Microsoft, whatever that means.

9.  Remember Borland, WordPerfect, Lotus? They all seemed mighty and indomitable in their day. They bet the house on one business model and lost. The big names of that era to have survived are Microsoft, IBM, and, to a lesser extent, Novell. All have changed radically, especially Novell. Microsoft can, and will, change equally radically to survive.

So, in ten years, we’ll look back with hindsight, and see Microsoft still as a mighty company (if not quite so dominant, perhaps, as it is now), and it will seem entirely natural that its core products are all open source. We will be slightly amused to recall that they weren’t always that way, because open source, is just the way software is developed, right?[7]

 Andrew Katz is a Partner at Moorcrofts LLP:


[1]    Lest you say: well, it’s the professional typists who count, and they all preferred WordPerfect, you need to know that (1) I was many years ago a temping secretary, and (2) I can type at 70 wpm. The preference was down to the familiarity of lock-in, something which made the transition to Word quite tricky, and something which Microsoft seems to have forgotten about in Office 2008 by radically reworking the user interface, and breaking a large proportion of the keyboard shortcuts. Which, as it happens, reached their pinnacle of speed, logic and simplicity in WordStar (and I can still remember most of them despite not having used WordStar for at least 20 years).

[2]    Remember Trumpet WinSock? It was one of the third-party components of the Windows networking stack which you needed to acquire before you were able to access the Internet.

[3]    See Helen Custer’s book Inside Windows NT, which Amazon informs me is now available new for $0.90, or used from $0.01.

[4]    Not GPL code, so far as I am aware

[5]    Microsoft Public License and Microsoft Reciprocal Licence

[6]    e.g. WiX (Windows installer XML) and ASP .NET Ajax control toolkit – see for many more examples.

[7]    As someone who’s much more interested in reading last year’s copy of Old Moore’s Almanack, than the current edition or old copies of Future Shock or the Popcorn Report (or even The Future of Law) so I can laugh at the daftness and chutzpah of their predictions, I realise that I am priming a fairly unstable petard. Still, this is just a bit of fun, isn’t it?