In many ways, not allowing protocols to be copyrighted is a good thing.
Webster's says that a protocol is a "set of conventions governing the treatment and esp. the formatting of data in an electronic communications system." Computer protocols, just like diplomatic protocols, are supposed to facilitate communication by making sure that everyone knows the rules. Thus protocols should be freely distributable so that a maximum number of people and computers are able to exchange information.
It becomes clear that Microsoft is perverting the very concept of a
protocol by not making theirs free and open. Now here's where it really gets
insidious: obviously their so-called "proprietary protocol" is useless if only
they know it. Even if they incorporated it into their own internal products it
would be relatively benign. But what Microsoft will undoubtedly do is leverage their market clout, and license this protocol to select developers for a hefty price.
Those companies which license Microsoft's mutant Kerberos protocol will be the only companies able to communicate with all
Kerberos-enabled servers everywhere. Open source Kerberos implementations, on the other hand, will only be able to communicate with each other. Microsoft's proprietary protocols are abusing the good will and open nature of the Open Source community. Microsoft builds on the free and open Kerberos protocol developed at MIT, but prevents everyone in turn from building on their own modifications.
In some ways, Microsoft's license agreement for the Kerberos derivative seems to be a specific snub to the open source community. As I read the agreement (see this article), anyone can agree to the license and then implement this protocol in a closed source product since a binary executable will not reveal the protocol! On the other hand, open source software necessarily reveals the protocol to everyone and thus violates the license agreement!
Here's my idea for a solution: Microsoft is specifically objecting to the verbatim distribution of a proprietary document on Slashdot. But what is this document about? It merely defines their version of the Kerberos protocol. Since, as I stated above, they can't copyright the actual protocol, why can't someone in the open source community summarize or rewrite this protocol and make their specification freely distributable? Just as KDE is free to emulate Microsoft Windows as long as they don't use any Microsoft code, aren't we free to implement their protocol without copying their document?
In the future, patents might be a way to protect protocol freedom. GNU is opposed to software patents, but a protocol itself really isn't software, and I see no reason why they could not create a patent version of the GPL. Just as GNU uses copyright law, in the form of the GPL, to protect software freedom, they might be able to use patent law to protect the freedom of protocols. I do not know the intricacies of patent law, however, and I am very curious to know if this is feasible.