I like beer.
I'm not talking about the insipid, flavorless swill that passes for mainstream American beer: Buttwiper, Swiller and their ilk. I'm not even talking about the Teutonic brews -- specifically lagers and Pilsners -- upon which that anemic American eyewash is based.
(Don't even get me started on the horrors of "lite" beers. That crap ain't even beer. It's just lightly-adulterated water -- and I, personally, prefer my water straight, no chaser.)
No, what I'm talking about is English-style beers -- especially ales, porters and stouts.
In other words, I like beers you can eat with a fork. Big, thick, meaty brews, heavy on the malt. With few exceptions, I prefer drinking beers that can pass the 60-Watt test. (Hold a standard pub pint glass of beer up to a naked 60-Watt light bulb. If you can see light through it, it flunks the test.)
You see, I drink beer for the taste.
A lot of people make that claim -- and then they turn around and guzzle a twelve-pack of Pabst or Schlitz or even
My secret is simple enough. I don't drink beer when I'm thirsty. I drink water, instead. And, when I drink, I drink only beer that actually tastes good, instead of the thin, astringent, flavorless slop that passes for beer in college rathskellers and sports bars across the USA.
Now, I realize that not everyone shares my taste in tipple. Some folks are teetotalers, for instance -- and I say, "More power to 'em." Others prefer hard liquor or mixed drinks.
And -- luckily for the Unix community -- Corel Corporation is among those who lean, instead, toward Wine.
Over the years, various parties have taken a stab at making Windows applications run under Unix. Religious objections on the part of OS purists aside, melding the stability and power of Unix with the user friendliness and desktop orientation of existing Windows applications makes a lot of sense. Unfortunately, those who have tackled the problem have, up to now, met with only limited and temporary success, at best.
In the Windows 3.x era, Sun probably came closest with Wabi (Windows Application Binary Interface). On Sparc machines, it consisted of both an emulator for Intel hardware and a translation layer for Windows API calls. Caldera Systems -- the current owner of DR DOS and one of the first ISVs to build a commercial product out of Linux -- licensed Wabi and did a Linux port, dropping the unnecessary Intel emulation software in the process. Sadly, neither version could run 32-bit apps and, when Windows 95 hit the market, both Sun and Caldera dropped further Wabi development like a dead rodent.
Meanwhile, other contenders stepped up to take a swing at the problem. They ranged from Onno Hovers' solo effort, WinFree -- which never made it out of alpha -- to commercial efforts like Mainsoft's MainWin, Bristol Technology's WindU, -- both of which are based on Microsoft's own Windows Interface Source Environment (WISE).
Hovers bailed when he realized that he'd bitten off more than he wanted to chew. MainWin and WindU work fine for developers who want to write cross-platform applications, but they don't do a thing for existing shrink-wrapped commercial apps. And WISE's other major licensee, Insignia Solutions, has sold its emulation product, SoftWindows to FWB Software LLC.
SoftWindows' principal market has always been Mac weenies with PC envy, rather than followers of Tux the Penguin, so the Linux world barely noticed its sale to FWB.
The other major player is the Santa Cruz Operation, whose independently-developed SCO Merge emulator -- currently at version 4.0 -- runs Windows 9.x apps. It's available free for non-commercial use. Unfortunately, Merge only runs under SCO's own OpenServer Release 5.0.4 or later, and, while OpenServer itself is also free for non-commercial use, the plain fact is that the world is fascinated by Linux, not SCO OpenServer.
Another swing by a commercial vendor, resulting in another clean miss. And all the while, much like Aesop's fabled tortoise, the open source community was slowly, doggedly pursuing its own solutions to the Windows-on-Unix problem.
The French Connection
Beginning in 1993, a theoretical discussion about Windows emulation on the comp.os.linux newsgroup evolved into an attempt to turn theory into reality. A programmer named Eric Youngdale got the ball rolling with a quick-and-dirty loader for Windows binaries he whipped up as example code.
That was enough to motivate a fellow named Bob Amstadt to organize what swiftly became the Wine project. Working with TK widgets, Amstadt set up a classical open source effort with himself in the Linus Torvalds role as ringmaster and benevolent dictator. However, it didn't take long before Amstadt tired of the job -- running an open source project is much harder and less rewarding work than most people imagine -- and turned it over to Frenchman Alexandre Julliard of the Swiss Ecole Polytechnic Federale de Lausanne, who remains in charge to this day.
Unlike Wabi, Wine never was an emulator. Since its first -- and still primary -- target platform was Linux, there was never any need for Wine to emulate Intel hardware, a la Wabi. Instead, from its very inception, Wine has consisted of a program loader, which loads and executes the Windows or DOS binary, and a library that translates Windows API calls into their Unix and X11 equivalents. That means that, in general, most Wine-compatible Windows programs run just as fast under Wine as they do on a native Windows machine.
It also means that -- unlike Wabi -- no actual license copy of Windows is necessary in order to run Windows programs under Wine.
In the beginning, most of Wine's developers were principally interested in running MSDOS and Windows 3.x games under Linux and they concentrated on incorporating those APIs into Wine. It was only much later, after the Wine effort weathered the transition to the 32-bit Windows 9.x era, that it attracted a group of programmers with more "grown up" uses in mind.
Which is where Corel entered the picture.
In the late 1990's, Corel Corporation began searching for a corporate strategy that would free it from Microsoft's dominance without forcing it to sacrifice its already-dwindling market share. With Macintosh sales plummeting like a rock in the post-Scully, pre-Jobs regime, putting its eggs in Apple's basket was clearly the wrong path. For all practical purposes, that left only Unix as a viable alternative.
After a brief and abortive attempt to craft a Java version of its WordPerfect Office Suite, Corel increasingly turned its attention to Linux. That made a certain amount of sense, since WordPerfect had always kept a toe in Unixian waters anyway, even before its disasterous acquisition by Novell and subsequent sale to Corel.
Beginning in 1998, Corel started devoting a substantial amount of its resources to Wine, as an adjunct to the development of both Corel Linux and WordPerfect Office 2000 for Linux. By December, 1999, according to Derek Burney, Executive Vice President of Engineering, about three dozen Corel programmers were working more-or-less full time on Wine. Their goal is to add sufficient functionality and stability to Wine to allow the WP Office 2000 for Linux and Windows platforms to share the same binaries -- and to move Wine from alpha code to a 1.0 release.
And that may just mark the beginning of the end of Microsoft's stranglehold over the desktop.
Not every Windows API will be supported in version 1.0 of Wine. There will still be a big hole in the communications arena, for instance -- and that's logical enough, since Corel doesn't make communications software. But it will include WordPerfect's trove of printer drivers and enough other functions that most Windows office productivity programs will run more-or-less seamlessly, regardless of who makes them.
Already, there is a host of Windows applications that run just dandy under Wine. Among them are Word 97, the BackOrifice remote control Trojan and a raft of games -- and they all run under Wine's .9x alpha code. And, when version 1.0 is released, the list is sure to grow substantially.
Wall Street has already embraced Linux in a big way -- it made a billionaire out of Red Hat's Bob Young, valued tiny, profitless VA Linux at $10 billion and rewards nearly every publicly-traded company that makes a Linux-related announcement with an immediate jump in its stock price. Clearly, investors see a change coming, soon -- and, with the Redmond behemoth under continuous legal fire as a monopolist, the Age of Linux may be even closer than it appears.
The prospect is enough to make me consider switching to Cabernet -- and it may even be enough to drive Bill Gates to drink.
(Copyright© 2000 by Thom Stark--all rights reserved)