From Visual Developer Magazine #47, December/January 1998


Lay Some Skin On Me, Linux!



There's a solid, cheap OS out there that beats the stuffing out of Windows 95 and stands fair against Windows NT. It's Linux—and all Linux lacks is skin.

Apple Computer just decided to kill its clone business. They obviously want 100% of the Mac pie, but it's a small pie that's shrinking daily, and it's really the wrong pie anyway. This is truly stupid. Now that IBM has thrown up its hands on OS/2, the Mac is the only Windows competitor still breathing, and ya gotta hold the mirror up real close. What Apple ought to do is get the hell out of the hardware business entirely, and port the Mac OS to Intel. The PowerPC just can't cut it-and Microsoft didn't get where it is by selling systems.

Ah, well. What all this is really about is how to keep Microsoft from owning everything. My contention remains what it's been for years: That Microsoft's competitors continue to do all the wrong things for reasons that no longer matter. Apple's pretty much out of the game. What's left?

Maybe we have to create a new competitor. And that's not as tough as you think. There's a solid, cheap OS out there that beats the stuffing out of Windows 95 and stands fair against Windows NT. It's Linux. Linux's problems are solely descended from the stupid Unix-hacker's ethos that if it doesn't make you bleed it's not healthy. It's a text-intensive OS that you can't place in the hands of anybody but another Unix enthusiast, and unless something intelligent happens soon it will close and lock the gate behind itself as a server platform and nothing more.

To become a desktop OS, Linux needs new skin. It occurs to me that with 13 years of Mac experience and eleven years of Windows experience behind us, we know what makes for an efficient GUI. The decisions today are far more how to implement than what to implement. (Don't even mention X Windows here—I boggle at the wrongheadedness of this architecture, reflecting as it does those university salad days when cycles were dear and bandwidth was cheap. Talk about backing the wrong horse…) We know what a reasonable set of windowing APIs looks like, and all the basic items—overlapping windows, icons, drop-down/pop-up menus, and so on—are in the public domain.

With Linux creator Linus Torvalds (according to Wired) in retirement, who's going to pull this off? The Unix community could, I guess, but they don't want to. So without any further introduction, here's my dream scenario:

Oracle buys one of the Linux companies—Red Hat, for example—to get a foothold in the Linux market, and puts a skunkworks together to design and implement the shell. In many respects that's the easy part. I'd suggest they work with Borland to get BC++ ported to Linux first and write the shell in that, because if the shell is completed, developers will need good tools to make porting existing GUI apps to Linux reasonably easy. (The Linux tools market is an empty box-and a big one.) Oracle then ports Personal Oracle to the new shell, along with any additional code necessary to make Linux an el primo multi-tier database client. Borland ports all their GUI IDEs to Linux as well—heck, if there were Delphi for Linux I might consider dumping NT. Finally, Oracle works with Sun to get a solid Java VM (virtual machine) in place for the Linux/shell combo. As a bonus, Netscape agrees to port Communicator as well—Linux lacks a good Web browser.

The idea, obviously, is to pull together all the viable non-Microsoft technologies onto one well-integrated platform in an Intel box. This will require a level of cooperation rarely seen in this business, but if Larry Ellison intends to remain in the Big Nerd Mud Wrestling Championships he's going to have to have his own operating system, and he's the only guy big enough to twist all those other arms into making it happen. All this talk of buying Apple (no! Not that!) and even the network computer (NC) pale in comparison to what a truly unleashed Linux could do for Oracle—and by implication, for anybody who really and truly doesn't want Microsoft to end up with 100% of the 21st century's trillion-dollar pie.