I wrote this article yesterday dinner time after reading the Groklaw article that I quote below, and I then ran out to fencing, forgetting to publish it when I got back. Since then there has been further coverage on Michael Meeks's blog and an article on Linux.com, 'Novell is not forking OpenOffice'.

Yes Novell is dating the beast, but that is not relevant here

When I first heard about the deal between Novell and Microsoft, my first thought was "Judas", and I would be the first to admit I don't understand why Novell agreed to the patent aspect to Novell-Microsoft deal. Apart from netting Novell a huge pile of money in the short term, it has lost Novell quite a bit of goodwill.

Frankly, I think the problem has been stoked by Steve Ballmer's big mouth, if he had not been sabre rattling then I think the community probably would have come to some kind of acceptance with the deal when the technical fruits started to emerge. What could have blown over in a week has continued because of Steve Ballmer's big mouth and his fondness for expressing how 'intellectual property' and software patents are going to bring world peace, feed the poor and save the world as we know it from cancer causing small companies and independent developers. Before, Ballmer was the cut-throat corporate pirate powering Microsoft along, but chair-throwing aside, his wackiness was somewhat held into check by Bill Gates who provided the calm voice of technical reason. Now Gates has left, it seems the balance has gone and the lunatic has taken over the asylum.

However, I certainly understand why, on a technical level, Novell would want to collaborate with Microsoft. This has already led to some interesting initiatives, Moonlight being the biggest example, but also it terms of making Linux work better in mixed environments and bringing OOXML support to Linux (we are after all going to have to read and write these things once in a while).

On a marketing level, it makes sense for Novell also. Imagine a Microsoft sales rep goes to a major company and says 'buy Windows servers', and the company replies 'sorry we are strictly a Unix shop and only use AIX' or 'we only buy SUN, go away', then the sales rep can go 'ah-ha, have you heard about this great product called Suse Linux? I can sell you a Suse Linux coupon for half the price of your current Unix contract'. As I often mention, we have not yet seriously began the Open Source VS Microsoft competition; Open Source and Microsoft are squeezing out the rest of the proprietary world first.

Novell is a big company, for the open source world at least, so saying that everything that everyone in Novell does must be a secret Microsoft agenda is a bit of a tinfoil hat conspiracy theory.

I read every article on Groklaw, it is one of my favourite sites. However, today's guest article by Charles Shulz is one of those tinfoil hat moments:

"And that's certainly an issue if Microsoft joins the game. Would that mean Michael's move was made in order to serve some corporate interests?"

I will try to explain the full context in a minute, but the short answer is no. Michael Meeks is not some Microsoft puppet. He is Linux/Free Software guy through and through. The problem has other roots. The Groklaw article and others are tying to claim that Michael Meeks has some secret agenda because they don't want to deal with his points directly. They are also ignoring SUN's motives, which seem at least as interesting as Michael's.

So let's start from the beginning.

StarOffice

In the 1980s, there was a German company called StarDivision that sold software for real computers like the Amstrad CPC and the Commodore 64. Then in the 1990s, they had an office suite, called StarOffice, for toy computers (PCs - DOS then Windows).

StarDivision was then bought by SUN in 1999:

"The number one reason why Sun bought StarDivision in 1999 was because, at the time, Sun had something approaching forty-two thousand employees. Pretty much every one of them had to have both a Unix workstation and a Windows laptop. And it was cheaper to go buy a company that could make a Solaris and Linux desktop productivity suite than it was to buy forty-two thousand licenses from Microsoft." (Simon Phipps, Sun, LUGradio Episode)

Apart from using it internally, SUN has experimented with various licensing schemes. Firstly, and most importantly perhaps, the Solaris Desktop now had an Office Suite, a big step forward. It also offered StarOffice 5.2 as a free download. It later offered to sell it as a vastly cheaper office suite than Microsoft Office, but as far as I could remember students and educational institutions could always get it for free.

Even when it was for sale it was never sold at a price that would actually make SUN any profit directly by selling to consumers, perhaps rightly, it was more about selling big companies SUN hardware and services and so on. If you did actually buy a licence, then you could install it on up to five computers of any operating system.

A couple of months ago, StarOffice went back to its roots as a free download as part of the 'Google Pack'.

OpenOffice

Not that long after buying StarOffice, SUN released the majority of the codebase under the Lesser General Public Licence under the name OpenOffice.org (the name OpenOffice was already trademarked).

To be honest, OpenOffice 1 was not that great. The Linux version especially sucked. It was usable, but just. As you ventured away from Writer, the other programs got particularly more hairy. However, it improved over the years and the 2 series was a dramatically better product.

By now I find OpenOffice, just as good (or bad) as Microsoft Office. I only use about 10% of the features of OpenOffice, so for me whether OpenOffice has 90% or 95% of the features of Microsoft Office has long become a moot point.

ooo-build

There are different ways to build programs. One way is to include everything inside the binary. This is how a lot of people build programs for Windows. This is not the Unix/Linux way of course. The Linux way is to have modular programs built on top of libraries that are already there in the system.

Another difference is that on Windows, you get one binary for everyone. On Linux you have distributors, who patch the package sources here and there, and generally work to make the package work better and be more integrated with the system as a whole.

What is best for the Windows platform is often not what is best for Linux. This is partly why Firefox is, in my opinion, so slow and ugly compared to Epiphany and Konqueror.

So Michael Meeks et al have been working on standardising these free/open source patches and submitting them to the upstream OpenOffice, while providing sources for Linux distributors to build from. If you read some of the blogs you would think that this was akin to Al-Qaeda.

Sauce for the goose

When Netscape open-sourced the Mozilla codebase, it ended up in the non- profit Mozilla foundation, a neutral place that volunteers and third-party companies can contribute code to. Another example is the non-profit Free Software Foundation; people contribute code and even assign the copyright of the code to it, confident that the FSF will keep it as free software/open source.

Another model is the Linux kernel. Here people retain sole copyright of their own code, but licence it under the GPL to Linus Torvalds et al who combine it all together. Linus et al see the idea of copyright assignment as contradictory, why is sharing under the GPL not enough to protect the work?

OpenOffice is different from most other Free/Open Source projects. You licence code under the LGPL, but then you also have to the assign copyright to one of the largest IT vendors in the world, who then have special rights to produce a proprietary product based on it. A proprietary product at least partially competing with the Open-source product you are trying to help out.

People contribute to projects for many reasons, some because they just enjoy it; while some people believe that free software is better ethically and/or technologically. These people are told that if you want your code to be in OpenOffice, then you have to contribute it to proprietary software as well. This is not the same as the BSD set-up where anyone can make proprietary offerings with it, here SUN has exclusive rights.

So the question is, how tenable is this situation in the long term?

If the root of the problem is actually StarOffice, then why doesn't SUN just kill it? It is not making money directly, and the original aim (to have a competing offering to expensive Microsoft Office) has been achieved in OpenOffice.

From a marketing perspective, OpenOffice is the better brand anyway, the employees currently trying (and failing) to sell StarOffice licences can offer per-seat support licences instead, think 'SUN OpenOffice Enterprise'. SUN has as much OpenOffice expertise as anyone, so it could make a credible offer.

However, far more important than per-seat direct sales, is the greater indirect value that OpenOffice can and does generate for SUN. Instead of the Microsoft Office model, think of Java. Anyway without dwelling on non- technical matters, there are lots of business reasons for SUN to support an open and free office suite; none of these require the existence of StarOffice too.

Sauce for the gander

Code is a literary product, like a book or a song, and the authors of code have a moral and legal right to choose what happens to it. Work for hire is of course owned by the company that pays for it, they have the moral and legal rights to choose what happens to it.

So we have had some people who have been willing for code to be used in OpenOffice under the LGPL but not in StarOffice under a proprietary licence. From the sidelines you can agree with it or not, you flame people like Michael in blogs, mailing lists and on Groklaw, however it is their right as authors to choose what happens to their code.

So if there is code that would improve OpenOffice and it is available to be used, then not including it because of StarOffice is in fact putting SUN before the open source community, StarOffice before OpenOffice. It is also a waste of time as the OpenOffice community has to go and rewrite it, just so SUN can include it in their proprietary software. Nuts.

An OpenOffice Foundation?

If OpenOffice was moved to a foundation, or the copyright was assigned to a third party non-profit organisation such the FSF, it would make things easier all round for open source volunteers.

So we have talked about SUN and the open source volunteers, but they are not the only stakeholders, the Linux distributors pay a number of people to work on OpenOffice, and there are other companies coming on board such as IBM. A foundation may help other companies come on board also, helping out an independent non-profit to produce open-source is a lot easier sell that helping out a giant IT company to produce proprietary software.

SUN did promise one when it released the code as open-source to start with, the initial press release said that SUN was going to set up a "new OpenOffice.org Foundation, which will initially be modeled on other successful open source projects and will consist of a project management committee, source code maintainers, and developers".

Does anyone have any idea where that went? They started to introduce some of the trappings of a foundation, e.g. a website and a mission statement but then it seems to have faded away. Come on everyone, let's finish the job!

Discuss this post - leave a comment