What is it specifically about the GPL that prevents from diving into the source code and identifying behaviour from that?

Google question Stu. Got an answer Chris, or is this just GPL FUD?

Matthew Holloway: "is this just GPL FUD"

Last I heard Uncertainty and Doubt were required for scientific reasoning. Fear is often quite healthy too. Why is it that when anyone raises a question about anything to do with OSS, it is shot down as FUD?

Instead, why not come up with some valid reasoning and evidence rather than just trotting out the tired old "FUD" excuse for closing down any rational question or discussion that you feel threatened by.

Person: "Oh look a purdy Tiger."
Guide: "Watch out, they can bite you, and I've been watching this one for a few weeks now. I'm not sure, but the way it acts, it just might be a man-eater."
Blissfully Ignorant Person: "What exactly is it about a Tiger that could hurt me?"
OSS Person: "Got an answer, Guide, or is this just Tiger FUD?"
Guide: "Huh?"

(Full disclosure: OSS People typically make more money if people ignore the dangers of Tigers.)

Stu: "What is it specifically about the GPL"

It's about the foundation of GPL. In other words, copy right law.

You can only use the copy right work to the extent the owner licenses you to. Since GPL does not specifically allow copying of features outside of GPL, this seems like a valid point.

Chris Auld: "reverse engineering by observation of behaviour isn't too risky from a legal view point"

Don't count on it. Copy right covers UI as well, and Apple and Xerox had a few goes at enforcing it in that area. I didn't really follow it but I think they ended up losing on technicalities. I don't think my lawyer would be as confident as you.

To avoid your GPL issue, perhaps you should use an arbitrary timing that someone came up with who had no contact with OpenOffice or your observations.

Wasn't <blink> started by Netscape as a non-standard extension to a standard? Interesting to see the wonders of <blink> being continued by IBM/Sun/OASIS. I have fond memories of entire websites blinking in Netscape because someone forgot to close a tag.

>It's about the foundation of GPL. In other words, copy right law.
>You can only use the copy right work to the extent the owner licenses you to. Since GPL does not >specifically allow copying of features outside of GPL, this seems like a valid point.

The GPL is a license for source code. It says nothing about what you can or cannot do with the ideas that you extract from that source code.

The intention of the GPL (or one of them) was to prevent a vendor from locking up an idea in a proprietary implementation. There is still nothing to prevent a vendor from implementing an idea in such a way; the benefit of the GPL is that there will still be a freely-available implementation of that idea that cannot be locked up.

The basic rules still apply: if you want to be squeaky clean about it, you do a clean-room spec. One person infers the behaviour and writes the specification from that. That person is then kept totally separate from the development team that implements the specification. Here we have an exchange of ideas, not implementations.

In other words, focus on process and technical details, not political or legal maneuvering.

Chris's original discussion above could also be couched in terms of "how easy (or hard) would it be to implement a good, interoperable implementation of OOXML?". That, I believe, would be more fraught with difficulty, working purely from the published specification.