Let the Sun-shine in (Oracle)

Contenuto non tradotto per le rilevanti dimensioni e il poco tempo. Descrivo perché prendo le difese (anche in senso tecnico) di Oracle nell'acquisizione di Sun.

First, you will pardon my cheap and obvious borrowing the lead tune from Hair.

In my previous blog entry I have briefly discussed that I intended to take actions to help the Sun/Oracle merger to be cleared by the EC Antitrust authorities. Indeed I have offered my dispassionate help as a Free Software and digital liberties advocate to the legal team assisting Oracle. They have gladly accepted my offer to help. It was discussed if I could also take the position of co-counsel to Oracle in the procedure, and so was decided.

This could be perhaps a surprise to casual readers. I take the opportunity to clarify what my motives are. It could be regarded as odd that I feel like I have to justify why I am accepting instructions that some of my colleagues would simply kill for. The point is that – unlike many – I am not a hired gun for whomever can afford to pay me. And I am quite fond of saying that I am in a position to refuse cases that are against my beliefs as a Free Software advocate.

Competition problems? Look at the alternatives!

Sun is a company that has lately been known for its huge contributions to Free Software. Indeed one could go as far as saying that amongst the big corporations Sun has been the most friendly to the community in terms of code released under Free Software licenses and general support. After cashing in perhaps too quickly a settlement on the same antitrust case against Microsoft that it had initiated in the first place (the one I have been deeply involved during the last five years), Sun has gained again good recognition and trust from the Community.

Sun has four important projects to Free Software and to a healthier and competitive environment in the software market. I am speaking about Openoffice.org, Java, Open Solaris, MySQL.

Sun is also financially bleeding, and its share price has been considerably dropping over the years. This, combined with its huge and easily cashable assets, made it a too obvious target for a takeover. The candidate for it are quite limited. Friendly takeovers could be made by IBM – who reportedly was involved in acquisition talks – and Oracle. Hostile takeovers could be made by Microsoft (subject to huge competition problems in most of the concerned markets) and by investment funds. The latter option is the one that frightens me most.

Sun has a huge patent portfolio too. Some of its patents have been pledged for instance to OASIS not to be asserted against any implementation of ODF. An investment fund would have no incentive to keep the company as a whole and, absent an overall industrial plan, the easiest cash-in move would be that of arranging an auction, one similar to that where luckily (the uninvited) OIN was able to recover the patents (dumped off by Microsoft) from the highest bidder. In other words, Sun or parts of it would become a litigation company, a pain similar to what SCO has been lately, but multiplied n-times.

We don't need to offer patent trolls any patents that are deeply embedded in core Free Software projects. We should remove the software patents from the equation, and we would be better off, but the threat is out there and ignoring it would do no good. Patent trolls, or technology investment companies – as they sometimes define themselves – could simply act on their own accord with the aim to force market players to pay through the nose or – much, much worse – could be a useful instrument to seed more FUD and to raise hurdles to the success of Free Software projects. To the best of my knowledge, Oracle is not asserting its patents against Free Software projects. But others could.

I think there is more than one reason to avoid all of that!

What is the sticking point in the investigation?

It's MySQL.

The Swedish database company has been acquired by Sun last year for a sum short of one billion euro. Not bad for a company whose best yearly turnover hardly reaches 50 million euros. MySQL and Oracle are databases. This is more or less where the similarity ends. Their products are significantly dissimilar from each other because they are at the opposite ends of the market. I happened to speak with people at Ingres recently, and they confirmed that, notwithstanding that they are also Free Software vendors with a business model similar to that of MySQL, they regard themelves as competing with Oracle and PostgreSQL, rather than with MySQL. but I don't want to argue that here. Complainants have sought to present a different view. The Commission decided that this point shall be further investigated and that's it.

MySQL is a very successful project in terms of deployment. I myself have no less than three MySQL instances in place and you are reading this blog from a LAMP installation – where "M" stands for "MySQL". It is very important that the project remains Free Software (it is licensed under a slightly modified GNU GPL V.2 on a dual licensing scheme) and a lively one. Yet I fail to see how a disbanded and broken-up Sun would do a better job at it than a company like Oracle.

Here we see the force and how dramatically Free Software departs from the proprietary setting. A product is very likely to survive despite the fortunes – or even against the will – of its founding company, because everybody is invited to take over or to fork. Forks are less frequent than one can think only because the simple threat of forking works in the direction of finding consensus among leading developers. Linux has never been forked. Samba has been forked (into Samba-TNG), but there is a constructive exchange between the two projects. Mambo has been forked into Joomla, and Joomla seems to be more successful than Mambo. More or less the samehappened to Xfree86. When the developers changed the licensing, a new project suddenly was born – X.org – and in six months it has swept away Xfree86 from the entirety of the GNU/Linux distributions out there.

Nothing substantially prevents the success of forks, apart from the success of the original product. We could say that the licensing reduces the barriers to entry in the market for that individual product, allowing third parties not only to take it as it is distributed by the project leaders, but to depart from it and make it – or parts of it – even an entirely different project. This is a terrific democracy-enforcer. The lack of democracy issue in the MySQL project has been latent for quite a long time, this is why a fork has already happened, into MariaDB, and if things get worse in the core development team of MySQL, it is easy to predict that the fork will have better chances to replace it – as it is drop-in replacement to the forked project. Nature abhors a vacuum.

If Oracle were hypothetically to bend the project away from competition in the high end or simply make it a stale project, it is clear to me that the declining fortunes of the original work would leave room (and disgruntled developers) to further the success of the fork(s).

Another misconception is that a Free Software project cannot succeed without a strong corporation as its proponent. All the evidence is to the opposite. Openoffice.org lives out a very lively and diverse community: while it has roots deep into Sun, many corporate users (including IBM) heavily contribute to its development. Linux prospers thanks to a wide community where no single company is relevant, leave alone dominant. Gnome and KDE the same. Again this is the strength of Free Software, that it permits many different development models, including the "bazaar".

Dual license is a moot problem

If I understand the point correctly, the objections on what I have just said is that – because MySQL is a dual-licensed product – a fork only relying on the GNU GPL licensing of the code would miss the revenues and investments coming from the proprietary licensing, which is out of reach of the fork.

This is absolutely frivolous, and it reflects a misconception of how the forces in the Free Software space work. It is not that a successful dual licensing enables a successful Free Software project, it is a successful Free Software project that permits to a dual licensing strategy to survive. The idea behind a dual licensing scheme is twofold. First you spread around a software application that becomes widely adopted, distributed, and where developers hurry up to develop for. And you offer it under the most restrictive copyleft license. Second, you offer a separate deal to all those who want to use the code but they are not willing to embed it into their products because they would be derivatives of the GPL'ed product, thus GPL themselves.

While dual licensing is often portrayed as the only successful business model for Free Software, this assertion is consistently disproved. And interestingly, this assertion comes from proprietary-intense (or proprietary-only) software makers. Its underlying assumption is that one can only make money out of selling licenses. So if somebody makes money out of Free Software, they must be offering proprietary licenses too. This misunderstanding results in frequent incorrect statements, such as – this is absolutely the most widespread myth – that Red Hat sells "commercial licenses" of its operating system.

This is the problem with Free Software: it changes any common-sense reference, it requires a change in the mindset. 2+2 = +∞ !

Dual licensing is hardly a viable model per se. MySQL Ab's revenues are pitiful in comparison with the market adoption of its project, and a considerable part of its turnover comes from other sources than licensing. MySQL is so good that it can be used via a lot of interoperability tools and via network without suffering big loss of performances, therefore the incentives to use a proprietary version come from the licensing issues of the very limited cases when a derivative product is to be distributed on proprietary terms. A proprietary standalone version of MySQL has no appeal compared to the Free Software licensed one. But in general, I know of very few cases of companies keeping a healthy dual licensed scheme (the only one I know for sure is Funambol). Dual licensing is very complicated to maintain, it requires copyright assignment, and this assignment is very hard to obtain from developers (this is at least my experience in drafting assignment agreements for clients).

Dual licensing is irrelevant in the success of the variety of Free Software business propositions, its fortunes will decline even more, and it can be seen at best as an interim setting to legal migrations from proprietary to Free Software.

Guess who's complaining

Oddly enough, the main complainants are two companies whose interest collide against the persistence of a competitive pressure coming from Sun. This is entirely appropriate, of course. Only, one could question how true the grounds are to complain of at least one of them, the one which competes directly with MySQL. It bugs me that they now potray themselves as advocates for the success of MySQL.

They are quite likely afraid that the Sun acquisition will reinvigorate MySQL to compete with them for instance in the SME's space, which is not Oracle’s strength. Another reason for them to fight against the Oracle/Sun merger is the resulting increased competitive pressure in the high-end part of the IT market. Especially that of large database installations for the cloud, a sector where it is no mystery Microsoft is trying to find a new expansion area, and which could be a very good sector where Oracle + Sun Solaris + Java could set the goalpost, while establishing a platform-diversity, under-the-hood interoperable ecosystem, with Oracle and Google and IBM on the upper hand, as well as benefiting the overall industry with huge Free Software spillovers.

The bitter end

Persistent delay as to the final clearance is sucking the breath from Sun, and rumors of people leaving are spreading. This cannot last. Continuing uncertainty damages Sun's developement teams. I have the pleasure to know a number of executives in Sun and – if perhaps they are not delighted by the prospect acquisition – they realize that it must be passed through as soon as possible, or the company will die. And with it, some of the good development teams that have considerably contributed to the success of Free Software.

It is time to take sides, and I know perfectly well where to stand.

Tipo di Entry: