OpenOffice.org 3 is here at last. It's an evolutionary change from 2, not a revolutionary one, and if deployed right it could be yet another nibble out of Microsoft's big Office cookie. The trick is how to get the word out.

For a long time, OO.o's depended on four things to gain traction: the fact that it's free; a small but significant set of corporate sponsors (Sun and IBM, mainly); frustrations with Microsoft Office -- mainly the price tag; and word of mouth. The last of those four, I'd put money down as being one of the most powerful: the word of someone you trust goes much further than advertising.

Each one of these four things has its limits, but the biggest is that, for the most part, many people simply don't know that OO.o exists at all. Or when they do, they're not sure how far it can go. Paula Rooney at ZDNet has urged IBM and Sun to not only market OO.o aggressively but offer free corporate-level training to those interested in switching.

I suspect those companies will only throw that much more money after OO.o if they're sure it will yield tangible benefits -- via support contracts, mainly, which is the most direct way to make OO.o a profitable enterprise without betraying its open source origins. But if they do invest money in advertising, one way to do this would be to pitch open source as a whole. The benefits of Linux as a departmental server (or Solaris as a high-end server), OpenOffice.org as a desktop application, and maybe a couple of other things -- all in one ad campaign.

Do this and you reap two benefits at once. You spell out the benefits of open source in plain language for a lot of different people, and you get the word out about a program that deserves to be used -- and put to the test -- by the broadest possible audience.

Final note: If you want to try out OO.o but don't want to formally install it, wait for the fine folks at PortableApps to deliver its no-install edition of OO.o 3. I've been using the beta and RC editions of OO.o 3 from them with no trouble. Right now the OO.o site is overloaded with downloaders ... which in itself might be a really good sign.