First, there needs to be a really good, robust Linux GUI front end to MySQL or PostgreSQL. By "really good" and "really robust" I mean that it needs to be just as good as Microsoft Access, and just as easy for end-users (and not geeks!) to use. There are some out there but two of the ones that pop up first on Google have had development suspended (MySQLGUI and MySQLFront). Another, MySQL Control Center, appears to be not really a database GUI but rather just what it says, a control center. Several others are out there, including SQuirreL SQL (Java-based), CocoaMySQL, kexi (part of K Office and perhaps the most promising at first glance), and finally PGAccess for PostgreSQL. I have not had a chance to evaluate these and I hope to do so in the future (in May, perhaps). Linux cannot succeed as a desktop OS without good tools, tools that are accessible to and usable by end-users as well as geeks.

Second, we really need OS X for Intel. OS X, Apple's new operating system for the Mac, is actually built on top of BSD Unix (a freeware version of the Unix OS) and there is no reason why OS X cannot be ported to the Intel architecture as BSD Unix runs just fine on Intel. Robert X. Cringely saw this a year ago; there is even an online petition asking Apple to do just that. Despite a lot of discussion (Ben Wilson and others) around last April, nothing has come of this so far, and Charles Haddad, writing in Business Week, thinks this would be a disastrous move for Apple. I obviously don't agree. [Late Postscript: John Dvork lays out a possible plan for Mac on Intel in his April 7 column, “How the MacIntel Will Change the Market ”]

Third, if IBM is REALLY committed to to Linux, they need to port Lotus Smartsuite to Linux. It's a real office suite which in most respects I believe is better that MS Office. One can only hope...