Do you have any information about how EU's initiative for investigation of Media Player started, since the whole case begun with Sun's complains on networking problems? Was it EU that noted this or someone else?

I was not involved at that stage, so I am no eyewitness of it. As far as I can tell from the documents, it was a "motu proprio" initiative of the Commission. In antitrust cases, the Commission has the right to initiate investigation also out of an informal complaint or even in case during its review of the market it finds infringements to the law.

What was Microsoft's strategy? How did the company defend itself in this case?

That is complicated. The general strategy appears to me, frankly, to be a damage control and delay one. They have hired possibly the best representation money can buy in Europe, people that have made the history of EC law. Luckily also on our side of the wall there was no shortage of brilliant and prepared minds. So the strategy was well served by a good display of tactical maneuvers to make the case as complex and costly as possible. The sheer volume of the paperwork is itself staggering: it takes almost a scaffold in my office. But I must also say that the opponent was reasonably fair in avoiding dirty tricks and ambushes.

The central point of the legal defence, at least in the interoperability case, was "intellectual property", meaning mainly trade secret and patents, but also copyright to some extent. Microsoft said that it invested a considerable amount of resources to invent new products and technology which are invaluable, and by releasing the information according to the requirements of the challenged decision would have allowed others to "clone" their products and in general to take a free ride on their inventions.

A quite spectacular defence was that about security. Basically it said that, unlike the Internet protocols, those keeping together a Microsoft work group network were so conceived that the all the servers acted as if they were a single distributed entity. In other words they were "tightly coupled", closely knitted together so that any intrusion from the outside, a drop-in replacement pretending to be a Microsoft Windows server could cause irreparable harm and all sort of nefarious problems. Besides, disclosing the specifications of their protocols would have required a hardening of the protocols, in order to make them resistant to malware attack or simply of badly designed third-party software which could have compromised the whole infrastructure.

But nothing was more shocking to me than hearing that they could not release the specifications because... they do not exist. They had to call back retired engineers or scroll through millions of lines of source code to find out what the heck they have done with the protocols.

Part of our value added in the case was to show how this all was nonsense. For the first part, our lead tune was "this stuff is not secret because valuable, but valuable because secret". Actually it is just a relatively thin proprietary layer of extensions to publicly available protocols and well understood techniques. At most, good engineering, but hardly any innovation if compared with existing implementations of the same protocols.

As the "tightly coupled" defence, the Samba guys were well positioned to rebut. Samba is quite a good piece of software, provided how it has worked out interoperability with Windows. So good that Microsoft itself used it as a reference implementation for the WSPP, the licensing scheme it has offered as a compliance to the remedies. And even without fully having understood all the idiosyncrasies of Microsoft's protocols it does not cause much disruption if used along with Windows servers. And I will not comment much further about the need to harden the protocols, as any reader could take the appropriate conclusion without me explaining how security through obfuscation is a false way of achieving security, and that malware makers surely do not need fully licensed specifications.