toby's Journal: Climbing out of the Microsoft coffin

In a 2003 article for Harvard Business Review, "IT Doesn't Matter," Nicholas Carr describes how "as IT's power and ubiquity has grown, its strategic importance has diminished," by analogy with infrastructural precedents such as electricity, rail, and telegraph. The principle is now even more evident than it was in 2003.

But the commodification of IT is only of advantage where companies are actually using best practices and getting benefit from the "sea of interoperability". Linux (and BSD) is the perfect example: a zero-cost, "set and forget" operating system layer that gets out of the way of higher ends*. Apache is another example: flexible, reliable and efficient, pushing concerns up to the application. Maturity is measured by problems solved - in this case a "stack" of components which are inspectable, fixable, well defined and well understood.

To take a favourite example, adhering to Microsoft is a net loss, since their agenda is explicitly to construct a proprietary "ghetto". Examples where this is a drag on efficiency are close at hand: SharePoint is not interoperable with standard browsers (e.g. Firefox); Exchange is not easily interoperable with standard clients (e.g. Thunderbird); you can only view Visio diagrams in Windows; DirectX (like all MS APIs) is designed to lock developers into Windows; and of course, everything is closed source and expensive (both adding to drag). The analogies are obvious: An electricity company that wants to sell you its 70 Volt appliances (or, for a closer analogy, they won't even tell you what the voltage is) while the rest of the world goes 110V; a rail company selling freight on its nonstandard gauge that unfortunately does not join all cities... No company that does not climb out of the MS box can claim full benefit of a commodified technological space.

The other great "drag" comes from reinvention: which I usually put down to "failure of research". In a maturing industry it is almost certain that any particular "need" (algorithm, language, protocol, format) has been solved before, usually by domain experts and to a level of sophistication or thoroughness that cannot be affordably duplicated (including by MS). This a natural benefit of the open source ecosystem.

Some developers evidently see themselves as Davy Crockett in a frontier world where you must whittle your own toothpicks and lumber, strike oil in your own backyard, and make clothing from animal pelts.

*update: A.N. Whitehead remarked that civilization advances by extending the number of important operations which we can perform without thinking about them (cited by Mark Dominus).

update(2): Sun's Jonathan Schwartz on 13 Nov 06, the day Java was GPL'd: A rising tide lifts all boats.

update(3): Duncan Cragg, 11 May 06, The "Imperative to Declarative Inversion": Open Data is OK!: Take a closed format such as Microsoft Word: such a document can only properly be unlocked after paying for Microsoft Office.

update(4): Matt Asay, 7 Dec 06, Weaning oneself from the proprietary nipple. It's strange to watch the convoluted dances we do to keep software proprietary. Why do vendors do it? Not because it provides one iota of value for customers. It doesn't. The opposite is true.