X
Tech

Microsoft breaks the perpetual licence covenant

Microsoft's decision to withdraw support for its own older file formats in the latest service pack for Office 2003 makes a mockery of the notion that users pay a single perpetual licence fee for conventional software. In reality, the costs never end.
Written by Phil Wainewright, Contributor

Many people fondly believe — in common with Talkback poster jim_d to my previous post — that when they pay a one-time fee to buy a 'perpetual' licence to use conventional software, then that's their only cost. But as I pointed out in my riposte, that fails to take into account the inevitable upgrade and support costs that come along with installed software. There's no escape from upgrades, because technology moves on, business needs evolve and security threats become ever-more sophisticated.

Vendors long ago realized that software is an ongoing rather than a one-time cost, but that gives them an increasingly big headache: their business model was founded on the same misconception that many users still make, of assuming a one-time 'perpetual' licence. This leaves them in the same sorry quandary faced in recent decades by asbestos manufacturers and tobacco companies, forced to bear a liability to former customers that they initially had no idea existed (and in some cases subsequently denied for years). Worse still, the liability has to be borne at the same time as a declining user base, making a proportionately larger dent in shrinking profits.

Microsoft's latest response to this quandary is breathtaking in its audacity. In what is surely the most pernicious act in modern software history, it has arbitrarily withdrawn support for older file formats of certain Office programs, including Word 6.0 and Word 97 for Windows, older versions of Excel and PowerPoint, and of some third-party programs such as Lotus Notes, Corel Quattro and Corel Draw. In doing so, it has broken an implicit covenant with its users that it will maintain backwards compatibility with earlier file formats. Zoli Erdos spells out the implications for users:

"Remember, this isn’t simply abandoning users still running pre-historic versions of software; we’re talking about data files here. You may run the latest release of all applications and still have no reason to touch old documents. After all, that’s what an archive is all about — you *know* your documents are there and will be accessible, should the need arise at any time in the future ... Microsoft just violated that trust, the very foundation of going paperless."

This is pernicious because Microsoft could very easily have chosen to bear the cost itself of continuing to support those file formats in Office 2003, as it has done to date. It claims the security threat has become too great, but it could have invested in building a mechanism to warn the user of the risks and offering to convert the file to a safer format. Instead it has decided it will no longer bear the cost of preserving access to its own file formats — formats which have become unsafe precisely because of its own eagerness to incorporate features such as programmable macros in the days before it discovered "trustworthy computing". It has left users bearing the cost of sifting through their archives manually converting the old files (a process described here).

This episode provides the most searing evidence imaginable to disprove the notion that a conventional perpetual software licence is a one-time cost, either for users or vendors. The need to preserve access to archived data (often with an audit trail, if it's business-critical information) imposes a duty and a cost that persists as long as the data has value.

As Zoli points out in his parting shot, this is a cost that on-demand vendors simply build-in to the honest, ongoing subscription fees they charge: "you don’t care about program versions anymore, just have access to your data. Anywhere, anytime." And the most reputable vendors will always allow you to extract that data any time to store it elsewhere, if that's what you want (so long as it's not someone else's data, that is).

Editorial standards