Enterprise Unix Roundup Sun's Open Source Pixie Dust
|Main||In Other News||Security Roundup||Tips of the Trade|
Last week, we made passing note of Sun's announcement that it would be releasing Solaris as some kind of open source software some time, somehow. We did so amid a flood of other announcements from the company of more pressing and immediate interest.
"Three years ago," we wrote, "this would have been like Nixon going to China. Now? [...] It's only a matter of time before the operating system doesn't offer anything users can't find in Linux or a BSD."
Proprietary software always faces that challenge when paired up against a free (as in no cost or source-available) alternative: Eventually, the matter becomes a John-Henry-like issue, as the core of hobbyist hackers are aided and abetted by companies happy to contribute a push along the way if it makes the free alternative better. A company with a proprietary product can close the source, but it can't hide the feature list, and eventually the open source version will accrete enough features to be competitive. Witness, for example, IBM's contributions to Linux, Apple's contributions to the khtml libraries that drive its Web browser and other chunks of HTML-based functionality in OS X, and Apache's assorted corporate benefactors. To survive, the proprietary vendor has a choice of succumbing or continuing to pile on features.
Of course, that's also the story between any two competing products. The difference being that open source or free software has the advantage of making stakeholders out of more than the core team around the project.
To listen to SCO or recent Linux nemesis Kenneth Brown of the Alexis de Tocqueville Institution, this is a matter of grave concern because, as Brown put it, the GNU Public License is "a leprosy," eating away at the bottom lines of hapless closed-source software companies.
For some reason, commercial endeavors, like Corel's release of WordPerfect, which dreams of being a competitor to Microsoft Word (and manages just that in some markets), is perfectly benign.
We're getting a little far afield here, but it's all in the service of pointing out that there's no pixie dust involved in opening the source to an application. As much as Brown, SCO, and the other assorted enemies of open source software try to paint Linux as a Destroyer of Value, they're ignoring the fact that they're simply dealing with a competitor put together in a fashion different from a software vendor selling a proprietary product.
And that reasoning cuts both ways.
Consider, for example, two of the more prominent and widely supported open source projects with origins in a proprietary product: Sun's own StarOffice/OpenOffice and the Mozilla Project.
The Mozilla Project came out of Netscape's Hail Mary pass to avoid being crushed under the wheels of the Microsoft juggernaut. To that extent it failed. The source code handed over to the open source community was an unusable mess, and when the developers who coalesced around the project finally obtained forward momentum, it was by starting over from scratch. Net result? Several years down the road, Mozilla is a very nice browser that did nothing to save Netscape and did very little to halt Internet Explorer's dominance. Any pixie dust anyone imagined working in that instance did not, and believe us when we say there was a certain intemperate type who promised Mozilla would be doing a lot more than "providing an alternative browser on a variety of platforms."
OpenOffice/StarOffice was a slightly different story. Sun, for starters, managed to keep better control of the process and hired experts to make sure expectations were better managed. Parts of the experiment have worked well to the extent that platforms that can't run Microsoft Office are in a much better position to be deployed these days. Other parts of it have not gone so well: We're still waiting for a native OS X version of OpenOffice ... not just one that runs on X11.
We don't want to seem like we're dismissing the value of either Mozilla or OpenOffice. Both are solid, usable products. Both have the value of keeping Unix users current in terms of desktop functionality. Both play well with other software and tend to observe the standards in their respective fields. But neither has proved especially destructive to its proprietary competitors, and neither has exactly set the world on fire.
Which brings us back to Solaris:
Were Sun to get around to opening the source code to Solaris some time before we all forget that it promised to do so, the company's own modus operandi with OpenOffice/StarOffice (i.e., maintain control, channel outsider interactions with the source through a carefully managed process, and maintain a dual license that guarantees the company maintains a measure of control over source code implementations) shows it wouldn't exactly be a free software hippie love-fest.
We also don't foresee the effect of the offering on the market being all that profound: Sun will likely keep the truly innovative stuff locked away; open source hackers (the benign kind) will likely poke around for a while, perhaps carrying off a tidbit or two to their "home" operating systems; and the Unixverse will continue to turn.
And it's not like Sun's in any hurry, anyhow. The company couldn't even tell curious reporters which license it plans to use (it didn't bother with the GPL for OpenOffice). That's probably just as well, since SCO (with which Sun holds some UNIX licenses) has already ruled out the GPL. As The Register noted:
"While the details of Sun's plan to open-source Solaris are not clear at this time, Sun has broader rights than any other Unix licensee," SCO marketing manager Marc Modersitzki told Computerworld. "However, they still have licence restrictions that would prevent them from contributing our licensed works wholesale to the GPL."
SCO's objections are pretty much by-the-by at this point, since Sun's not making any public comment on how it intends to proceed. Whatever accommodations it comes to with SCO, and however it moves ahead with the opening of Solaris, it's clear that Sun will continue to sell a services stack on top of an operating system it's valuing less and less.