Keeping Data in Its Place

Few things are more dangerous than data that doesn't know its limitations.

Few things are more dangerous than data that doesnt know its limitations. When data runs around with no idea of how it should (or shouldnt) be used, software stagnates and business models die. n Application developers know the hazards of letting data roam the streets. When any piece of an application can use any of that applications data, developers find themselves hamstrung by tentative decisions. Modules get tangled in mutual dependencies; software, in the words of "The Unix-Haters Handbook," "ceases to be soft."

Object-oriented software tools wrap data in access restrictions, allowing only the proper operations to take place. Developers retain their freedom to change internal representations, as long as an object still gives the expected answer to any question. Its a good thing, though the OO label is often used without being earned.

Content creators are trying to achieve at least as much success in associating usage rights with data, using technologies such as XRML (Extensible Rights Markup Language). Pioneered at Xeroxs Palo Alto Research Center, now championed by the companys ContentGuard spinoff (partially funded by Microsoft), XRML offers new hope for digital rights management: Its favorable prospects earned it a place this month on Technology Reviews list of 10 emerging technologies "that will change the world."

Some software writers chafe at the restrictions of object-oriented methods; some content users claim that copyright is theft, and that information wants to be free. Both of these groups look for ways to tunnel through any protective barrier around the data that they demand.

Its in all of our interests, though, to protect and reward the creative process—whether the product is software, or any other work of art.