Duffbert and Brian both commented yesterday about Dan Farber's report of an encounter between Steve Gillmor and Ray Ozzie. Steve asked Ray if Lotus Notes is dead. Ray gave good reasons why it isn't. Dan came to the seemingly odd conclusion that Notes is a "legacy application"... but maybe it's not so odd.
There's an alternative and growing point of view — growing since the "Dot Com" bubble burst in 1999 — that legacy systems are simply (and only) computer systems that are both installed and working. In other words, the term is not at all pejorative — quite the opposite. Perhaps the term "legacy" is only an effort by computer industry salesmen to generate artificial churn in order to encourage purchase of unneeded technology.
IT analysts estimate that the cost to replace business logic is about five times that of reuse, and that's not counting the risks involved in wholesale replacement. Shareholders and managers are increasingly asking, "Why are we spending so much money on new technology with so little to show for it?"
Businesses and governments are also recoiling at well-publicized system failures and security breaches that all too commonly arrive with new software — failures which are utterly catastrophic in many cases.
Increasingly the IT industry is responding to these understandable business concerns. "Legacy modernization" and "legacy transformation" are now popular terms, and they mean reusing and refactoring existing, core business logic by providing new user interfaces (typically Web interfaces) and service-enabled access (e.g. through Web services)
The reexamination of attitudes toward legacy systems is also inviting more reflection on what makes legacy systems as durable as they are. Technologists are relearning the fact that sound architecture, practiced up front, helps businesses avoid costly and risky rewrites in the first place. The most common legacy systems tend to be those which embraced well-known IT architectural principles, with careful planning and strict methodology during implementation. Poorly designed systems often don't last.
It seems to me that Notes and Domino are a poster-child for this view of legacy. It's installed for more than a hundred million of users. It works. It would be very costly to replace, and it's doubtful that there would be any business benefit to doing so. It's secure. IBM has been continually modernizing it, transforming it to work with new technologies, and keeping it working on the latest hardware and operating systems. It's a sound architecture.
Look at the mainframe. Look at a lot of the software that runs on mainframes. They've been classed as a legacy system for as long as many of us have been in this business, and they're still here, aren't they? They're still storing and managing most of the world's data. They're still being enhanced and companies are still investing in them.
You have to be successful -- very successful -- before you can be legacy. When you are legacy, you're going to take a lot of pot-shots from people who don't understand you and from people who want their mark by making change for the sake of change, but history shows that you're going to still be there after most of your challengers have come and gone.
I don't think I have any problem at all with Dan Faber's conclusion.