Computerworld

Will StarOffice on Linux kill Microsoft Office?

by now you surely have heard that Sun Microsystems

Inc. plans to release StarOffice 6.0 this fall as open source under the GNU General Public License. By coincidence, I had already returned to StarOffice 5.2 as my primary productivity suite after flirting with alternatives such as Applixware and Microsoft Corp.'s WordPerfect Office 2000.

Some readers have already accused me of being overly fickle, since I seem to switch productivity applications almost as often as Al Gore changes campaign strategies. Well, they're right. I've been leaping from this suite to that because I've never been completely happy with any Linux productivity applications, although I came closest to nirvana with WordPerfect 8 as my word processor.

By the way, I just did a quick survey of the files I have in my document directory tree, which should give you an idea of how much I've been jumping around from product to product. Before I break it down, I should mention that there are two misleading numbers. I did a lot more work with Applix Words in the past, but I backed up those files and removed them a while ago.

Unfortunately, I don't recall how many there were. On the other hand, I haven't gotten around to removing one branch from my tree that has a lot of old Microsoft Word documents. I haven't used Microsoft Word in ages, so the 90 Word files are pretty ancient.

Anyway, here's the current snapshot:

10 Applix Words files

31 Rich Text Format files

42 plain text files

90 Microsoft Word files

134 StarOffice files

243 WordPerfect files

As much as I like WordPerfect, WordPerfect 2000 for Linux is just too bloated, unstable, and unpredictable for me. For example, woe to the person who attempts to save a WordPerfect 2000 document as a PDF file. On my system the suite crashes and refuses to start again until I recover the configuration files from a backup or regenerate them from scratch. If you're having better luck with WordPerfect, more power to you because it is a terrific suite if you can keep it running smoothly.

I'm much happier with the stability and speed of Applixware. I would make it my default productivity suite but for a couple of factors, the most important of which has nothing whatsoever to do with the quality of Applixware: All but one personal information managers for Linux stink to high heaven -- the one that comes with StarOffice. The StarOffice PIM has a very nice interface, and it works great and synchronizes beautifully with my PalmPilot. That and only that drove me back to using StarOffice. I'm still unhappy with how RAM-hungry and slow StarOffice can be, but since I have to run a huge suite to get to the PIM, there's no point in starting up another word processor like Applix Words, since the StarOffice word processor is already there.

StarOffice under the GPL

That brings me to a number of details beyond the switch to GPL that make Sun's announcement interesting. The most intriguing fact is that Sun plans to split the office suite into individual application components. There are several reasons why Sun needs to do that. On the surface, it will make StarOffice more usable. StarOffice is currently a monolithic suite that takes longer to launch than it takes Microsoft to release a security patch. As I said, most of the time I only use two pieces of the suite, the word processor and the personal information manager. It would be a great relief to be able to launch only the applications I use, especially if that translates into faster launch times and less memory consumption. Who knows? If I could launch the StarOffice PIM separately, I might even go back to using the Applixware word processor, although I'd be perfectly happy with the StarOffice word processor if it just didn't eat up so much RAM.

Sun also wants to give the StarOffice components a Bonobo compatibility layer (Bonobo is the CORBA-compliant object model used by the GNOME desktop). I was a bit ambivalent about that decision at first. On the one hand, I am somewhat sour on GNOME of late and would have preferred to have seen StarOffice integrate more tightly with KDE. But as much as I prefer KDE to GNOME, I have to admit that the KDE team could turn out to be shortsighted with respect to its component strategy. On the down side, CORBA is sluggish and complex, and it is largely responsible for the lack of responsiveness of some GNOME tools. But I still believe the GNOME developers got it right in choosing CORBA, at least in the long run. Computers will continue to get faster, which will eventually make the speed a nonissue. And developers will continue to write better tools for creating and integrating CORBA components, making those components easier to use and manage.

It probably doesn't matter anyway. Greater interoperability between KDE and GNOME components is supposedly already in the works. Sun's choice of Bonobo will simply increase the pressure on KDE to make its component architecture work seamlessly with Bonobo.

More important to Sun is the fact that splitting up StarOffice into components makes the suite more likely to become ubiquitous. Sun's idea is to encourage developers to drop the StarOffice word processor component into their applications instead of writing a word processor or editor module.

There's no guarantee the open source community or commercial developers will buy into that idea, but I'm betting they will. The Bonobo layer should make the components relatively easy to reuse. There are no licensing fees or restrictions (other than the GPL requirement to release your modifications) to discourage developers from using the StarOffice components. Sun is establishing a foundation of XML-based open file formats for its documents, so developers and users don't need to be afraid they'll be locked into a proprietary data format when they adopt StarOffice or applications that use StarOffice components.

And Sun doesn't seem to be asserting exclusive control over any standards, rights or specifications that would give it implicit control over developers who adopt StarOffice components. Put simply, Sun offers the convenience Windows developers already enjoy when they reuse Word or Internet Explorer as part of their applications. But there are two dangers inherent in reusing Microsoft components. First, each new Microsoft service pack routinely replaces existing components with new versions that break the applications that depend on them.

Second, there is the well-known danger of developing anything for a Microsoft operating system, whether you use Microsoft components or not. If Microsoft decides any given application constitutes competition, its developer might as well file for bankruptcy.

The key here will be momentum. If Sun can convince enough people to use StarOffice components in the first year or two after the new version is released, it will behoove most developers to jump on the bandwagon just to ensure compatibility between their applications and what has become the mainstream.

Indeed, if enough developers jump on the bandwagon, Sun could succeed in cutting off Microsoft's air supply (to coin a phrase). As friendly as Microsoft Office may be (which depends on the components you use and how you use them), why would anyone pay to use Microsoft Office given its many disadvantages? In the first place, Microsoft designs Office to keep you on a costly upgrade treadmill. StarOffice is free, as are any upgrades. Microsoft controls its own document formats and Office APIs. The StarOffice document formats and APIs will be completely open. Microsoft designs its Office suite with one priority: to maintain its monopoly. Even if StarOffice ever attains a monopoly, Sun cannot manipulate the market with an open source product to maintain that monopoly.

Anyone is free to take the source code and turn it into a product that serves customers better.

All in all, I think Sun has it in the bag. I wouldn't be surprised if we see a nearly universal conversion within five years to StarOffice components as the basis for the dominant productivity solutions. What do you think?