A WSJ story talks about Microsoft’s bets on Longhorn and connects it to Cairo, the OS project which failed primarily because of the refusal of various parties to change user programs and file formats. Longhorn aims to do just that, providing among other things, a seamless way to search for information.
Cairo was once touted by Microsoft Chairman Bill Gates as a radical “paradigm shift” in computing, mainly for its lofty goal of overhauling and unifying Windows’ file systems. For everyday computer users, the benefits could include being able to seamlessly search through e-mail messages, Microsoft Word files or even Internet pages in the same way, instead of using one of the many separate file systems that exist today.
That goal, described by many people close to Microsoft as a sort of “holy grail” for Mr. Gates, didn’t work. Now, Cairo looks more like an embarrassing footnote in Microsoft’s product history.
One thorny problem for Mr. Allchin and his engineers is that building a brand-new file system — and replacing the different systems for storing things like Word documents and e-mail messages — could require the expensive and time-consuming rewriting of many of the software programs that run on top of Windows, including those made by outside companies. Those critical programs run the gamut from antivirus software to Adobe Systems Inc.’s popular Acrobat reader to Microsoft’s own Office suite of programs.
“If they change the file system, they are going to break programs,” says Michael Cherry, a former Microsoft program manager who is now an analyst for the firm Directions on Microsoft in Kirkland, Wash. Resistance from the Office group helped sink previous Microsoft efforts to unify file systems, and Microsoft should expect “a lot of resistance from Office” again, says former Microsoft executive Usama Fayyad, now president and chief executive of data-mining company digiMine Inc. He thinks a new file system is “a nice vision, [but] it’s not a reality, and it won’t be a reality in the short term.”
To make matters more complicated, Longhorn will include a new set of “application-program interfaces,” or APIs, that provide a recipe to make software that works with the new operating system. Those APIs are expected to improve things such as 3-D graphics in Windows, but they could also require changes in some existing programs for them to work properly.
Longhorn is scheduled for release in 2004 or 2005.
This is where I see an interesting opportunity for Linux and server-centric computing. Some of the innovations, especially with the file system, can be applied to the new set of users, who have little legacy. Already, all my files, emails, bookmarks, sites I am browsing, etc. are all there on the server. What is needed is a way to tie them all together. This is of course just one aspect of what an OS does. But by taking some ideas from the past (and the future), serve-centric computing with Linux can actually provide a better computing environment.
It is the same logic which we are trying to apply to the front-end with the Digital Dashboard project, by providing an integrated view of all the events and information flows that a user needs to see.
What is key here is the target audience: they are the new markets, new users. We have an opportunity to rethink the computing platform for the next 500 million users and apply these innovations without worry about the past for these users.
It is, in some ways, the same thinking Reliance Infocomm is applying to the telecom market in India – by providing a very rich set of features and services in the coming IndiaMobile service with CDMA 1x (which is for all pratical purposes a variant of 3G) to users who have never really experienced mobile telephony or the Internet, and all this at an affordable price point.
Use technology to help new users leapfrog – that should be how we need to think.