It has been months now and I'm still receiving letters about my first rant. The basic thrust of the rant is that Linux developers should be focusing more on innovation than on mimicking what is already on Windows. I stated what I thought were good arguments, and I had many more that wouldn't fit into the space available for my column.

Most readers applauded that column. Some disagreed, and they had some pretty good arguments, too. The best argument revolved around desktop productivity software. They argue that Linux office suites must mimic Microsoft Office to some degree, but mostly with respect to document format. It is undeniable that most business desktop users are running Microsoft Office. It will be impossible to woo these people away from Windows and Microsoft Office unless a Linux suite can make the transition away from Microsoft Office an easy one. You can't do that unless the Linux office suite can read and write all those legacy documents seamlessly.

Put another way, the only way a Linux office suite can beat Microsoft Office is to (essentially) be Microsoft Office, at least until the the Linux office suite has gained enough market share and momentum to unseat Microsoft Office. This is an excellent argument, though not a perfect one.

Let's go over a little history. Once upon a time, WordPerfect owned the word processor market and Lotus 1-2-3 owned the spreadsheet market. They had no credible competition. Then one day, we woke up, and Microsoft Office had taken over the world. Okay, so it didn't take a single day, but it sure felt like it at the time.

How did Microsoft accomplish this mighty feat? Did Microsoft make its products so much like WordPerfect and Lotus 1-2-3 that the transition was seamless? No. Microsoft added lots of compatibility features in its products to make the transition from WordPerfect and Lotus 1-2-3 less painful, but the transition was in now way seamless. Word never worked like WordPerfect (to the dismay of many), and Excel never worked like Lotus 1-2-3. Microsoft provided some similarities, but OpenOffice.org is far more similar in look and feel to Microsoft Office than Word and Excel were to WordPerfect and Lotus 1-2-3.

I can sum up the way Microsoft beat the competition in just a few words. Microsoft won the market by leveraging its virtual monopoly on the desktop. Microsoft could co-develop its office products and Windows to play nicely. The competition couldn't. This is one reason why WordPerfect and Lotus 1-2-3 couldn't make the transition to Windows easily. The developers weren't privy to the APIs (and their quirks) as early as Microsoft, and the WordPerfect and Lotus developers couldn't tweak the Windows API in order to make their products run better. There were other factors, too. For example, Microsoft sat on enough cash that it could undercut the competition until it won the market.

No pain, no gain

Here's the moral of the story. Microsoft did not provide 100% backward compatibility with existing WordPerfect and Lotus 1-2-3 documents. Microsoft did not provide perfect look and work-alikes to WordPerfect and Lotus 1-2-3. Granted, Microsoft had the edge in integration and (to a degree) stability, because Microsoft owned the API. But overall, Microsoft did not win the market by providing applications that were superior to WordPerfect and Lotus 1-2-3. I would argue that the WordPerfect of 10 years ago is in many ways superior to Microsoft Word today, but that's neither here nor there. The point is that to one extent or another, people were willing to "downgrade" from some of features of their favorite applications, retrain their users, and suffer the transitional difficulties in order to migrate to Microsoft Office.

In short, history shows that people do suffer transitions. They do retrain. They do migrate from one document format to another. While pain may not be necessary, people are willing to suffer pain in order to get enough gain.

Therefore no Linux office suite has to "be" Microsoft Office for any period of time in order to displace Microsoft Office in the market.

What have we learned?

So what will it take for a Linux office suite to displace Microsoft Office? One or both of the following:

1. Linux has to gain a virtual monopoly on the desktop
2. The Linux office suite has to be compelling enough to motivate people to migrate away from Microsoft Office

Linux has arrived on the desktop. There's no reason why anyone needs to run Windows. But let's face it. Linux won't monopolize the desktop anytime soon. So there's no way to leverage a desktop monopoly to get people to switch away from Microsoft Office.

However, there are ways to make an office suite compelling enough to draw people away. Here are those factors that I find compelling (your mileage may vary):

1. The office suite runs equally well on Windows and Linux to ease a transition
2. The office suite looks and functions (UI, reads/writes documents) enough like Microsoft Office to take most of the pain out of migration
3. The office suite avoids or corrects all the design errors in Microsoft Office
4. The office suite provides compelling innovative features that you can't find in Microsoft Office

We already have a couple of good office suites that satisfy items 1 and 2. OpenOffice.org runs as well on Windows as it does Linux (some might say it runs as badly on both platforms, but that's a matter of tolerance and taste). EIOffice runs quite well on both Windows and Linux. Both products read Microsoft Office documents well enough to make a transition possible with only minimal pain.

That leaves us with points 3 and 4. These are the points that motivated the original rant that drew both praise and criticism. The most compelling feature of OpenOffice.org is that it is free. Beyond that, OpenOffice.org duplicates many of the brain-dead design errors you find in Microsoft Office. This is not the way to draw people away from Microsoft Office. You draw people away from Microsoft Office by eliminating those features that Microsoft Office users find frustrating, and providing features that Microsoft Office users wish they had.

It is easier to point to what EIOffice did right than pinpoint the flaws in OpenOffice.org. EIOffice made it ridiculously easy to create live links between documents and document types. If you want to plug the result of a spreadsheet calculation into a document or presentation, all you have to do is "copy" the spreadsheet cell and "paste as link" into the document or presentation. Once you create the link, it is nearly impossible to break it. You can cut the formula in the spreadsheet and paste it to another location. You can add rows or columns so that it moves the location of the formula. It doesn't matter. You can change the data in the spreadsheet such that the results of the formula change. EIOffice maintains the link and instantly updates the results everywhere.

At the time I wrote the rant (and as far as I know this is still true today) there is no easy way to create live links in OpenOffice.org or Microsoft Office. When you do create live links, they aren't nearly as robust.

Something you can't live without

The key here is that once you start using the live links in EIOffice, the linking becomes a feature you can't live without. This alone would provide a compelling enough reason to migrate away from Microsoft Office. Unfortunately, Evermore Software (the company that produces EIOffice) isn't pushing its product well enough to give me hope that enough people will try EIOffice, get hooked, and adopt it. And there's the issue that EIOffice, while a bargain by any standards, isn't free. And it isn't open source. These may be reasons why EIOffice will never replace Microsoft Office, but these issues do not detract from my point - that EIOffice had the right idea by offering innovative ideas and avoiding the problems inherent in the Microsoft Office design.

Ironically, EIOffice is an almost perfect duplicate of Microsoft Office in terms of the user interface. So Evermore Software is focusing a lot on taking the pain out of migration. But I still think its most compelling feature is its approach to live linking.

That is why I stand by my original assertion. Linux developers should be working on imitaing what Evermore Software did. Sure, minimalize the pain of migration. But focus on avoiding the pitfalls of Microsoft office, and on providing the compelling features that make people want to use the Linux office suite instead of Microsoft Office. Instead of prompting people to say "It's free, and it's good enough", make people say "It blows away Microsoft Office and, oh, by the way, it's free, too."

Do that, and you'll find that people will migrate despite the minor incompatibilities in user interface and file formats.