Joel Spolsky's latest is a valuable SITREP on the ongoing Platform Wars between Microsoft and everyone else. If you are a software developer or entrepreneur, read the whole thing. For others, his thesis is that Microsoft itself has destabilized developer support for the Windows APIs, heart of the platform franchise. And it has done so at a time when the Web is seen as an alternative to many 'rich client' apps, and investors are scared away from financing PC-borne software. The result is a weakening of the Windows franchise, and a shift of the platform war towards the server platform.
Though I've not had to suffer through the travails of COM and .NET as has Joel, I've seen the effects. He's dead on about the investor sentiment. The purpose of this post is to add a little historical perspective that might suggest where things go next.
Once upon a time computers were - well - computers. That is, machines mostly used for calculation. Those of us of an age carried card decks as supplicants to the glass temples of the mainframes. Later on, TTYs and 3270s replaced the card decks for data entry, and many could use the same machine at once. This was called timesharing. One machine, many users, all the same interface.
Then came the PC. The computer became personal, and as such its nature changed. Whether used for calculation (Visicalc) or writing (Wordstar, Word), it became a personal tool, an extension of the user. The computer as tool was best when it got out of the way, and presented the task in its own terms as much as possible. Hence, the rise of direct manipulation interfaces, with Mac and later Windows as their embodiment. This is the type of software which Joel calls 'rich clients'.
The advent of 'multimedia' and CD-ROMs gave the first hint of the next evolution of computers - into communications devices. This hit its full stride with the coming of the commercial Internet, the Web, and the social acceptance of e-mail. The relative asymmetry between creators and readers put more emphasis on predictability in the interface, rather than power as a tool. 'Computers as media' are following previous new media, by evolving into genre of usage which provide simple, recognizable patterns for messages (you're looking at one of them). Although some of the data elements used are called 'rich media', the computer itself is just a conduit for that data delivery.
The power of Microsoft rose with the rich, direct manipulation, GUI software of the PC age. But 'timesharing' as concept never died during that era. It faded into the background, and remained a latent competitor. How have the two styles fared as the battle moved on from computer as calculator or tool, to communications device?
It should be clear that timesharing has been reborn in new clothes. Now we call it 'salesforce.com' and 'Google'. The 3270 of old is now an HTML or XML browser. This interface has absorbed most of the new values of the communications function: Web, blogs, increasingly e-mail. The genre of the Internet are evolving to fit the limits of the browser, for better or worse. The business models are subscription, fee for service, or advertising. With the foundational layers of server software increasingly commoditized, investment flows towards value added services, and once again toward complex calculations - analytics, ad targeting, etc.
The rich client is struggling. In retrospect, the attempt to introduce 3D to the browsing interface in the mid-90s looks like a prelude to its descent. In the end, users wanted simplified access to information more than a 'richer' tool that introduced its own set of problems. The same fate has met most attempts to create 'rich' client dependencies, often in the service of advertisers. The only domain in which the full client platform is exploited and pushed is gaming. A growing business, but not one on which to balance the whole Windows franchise.
Elsewhere, the client side is stagnant and investment stays away. Though Microsoft apparently recognizes many of the information management problems that the combination of PC and Internet has created for users, its solution - Longhorn - keeps slipping out in time and losing relevance. GMail and other net-borne palliatives are already arriving.
I am not forecasting the death of the PC. Larry Ellison already tried that once. It will be with us for a long time to come, due to its overwhelming scale economies. It will still run all of the tools that we use to create information: Word, Photoshop, Powerpoint and the rest. It will be a terminal for the new timesharing, and a conduit for the Internet as medium. But it is no longer the location of new investment and innovation, and Microsoft's ability to extract revenue, margins, and strategic advantage must fade accordingly. Joel's right: The heart of the Windows franchise is rotting.