In many ways paper data storage is superior. You don't need a machine to read it. You can modify it with your choice of a wide array of portable, inexpensive devices that need no power to operate. It's lightweight, flexible, and portable. It's universal -- anyone who knows how to read can extract the data stored on it. You can't say that about digital storage. Even if the storage medium survives any length of time, will you still have the hardware and software to read it?
So in a fundamental way, computers are a huge step backwards. Digital data storage is like the secret texts of the temple. Ordinary mortals are not allowed access. Only select individuals with the proper magical devices and incantations are allowed to see it.
A year ago IBM's CEO Sam Palmisano issued an ambitious challenge to his company: "...all of IBM to move to a Linux based desktop by the end of 2005." They still have a year to go, and already have an impressive number (over 15,000) of Linux desktops. But it seems there are some pretty large bumps in the road, such as Web applications that only work with Internet Explorer and ActiveX; no Linux client for Lotus Notes; and support staff who don't know Linux. Even if IE weren't the most malware-friendly browser on the planet, what is gained by excluding users on Unix, Linux, MacOS, and other platforms? Why use any server that only serves a single client? Servers should serve all clients. That's why we have standards for things like TCP/IP, POP3, IMAP, Ethernet, FTP, HTTP, XML, wireless, and so forth.
See, we grumpy oldtimers rail and crab about standards for a reason. Standards are what give us the freedom to choose from a variety of vendors, and to choose the technologies that best suit our needs. Data files should be not stored in closed proprietary formats either: text documents, spreadsheets, presentations, databases....none of it. I can't think of a single justification for holding your data hostage to a single vendor, but I can think of plenty of risks: vendor goes out of business, no competitive pricing, vendor does not deliver good service and fixes, difficult data recovery from failures...yeah, we've all been there. Everyone who is nursing some antique machine and praying that nothing breaks because you can't get parts for it anymore, because it runs an important but obsolete program that uses some weirdo data format that no one alive understands, raise your hand.
Future-Proofing Your Data
A good first step in migrating to a nice new open platform is to start at the server level. Remember, servers should serve all clients.
Then start looking for desktop applications that are both open data formats and cross-platform, like Mozilla, Firefox, OpenOffice, Thunderbird. The more you can standardize without excluding client platforms the easier your life will be.
That's the easy stuff. Then it's time to start looking at all those nice in-house custom apps, and the less-common, more specialized applications. Apologizing in advance for the grandiose pronouncement – our most important job for the new millennium is future-proofing our data. As IBM is finding out, the deeper the closed, proprietary, non-standard hole you are in, the longer it takes to dig out. But the alternative is being forever beholden to vendors who see us as fat, compliant profit centers, rather than customers.
Carla Schroder is a self-taught Linux and Windows sysadmin, who laid hands on her first computer around her 37th birthday. She is the author of the Linux Cookbook for O'Reilly, and writes Linux howtos for several computer publications. Carla is living proof that you're never too old to try something new, computers are a heck of a lot of fun, and anyone can learn to do anything. Visit tuxcomputing.com for more Carla stuff.