Uh oh, you're on a paranoid trip now. Probably no matter what I say, you'll presume it will be some kind of attack against you.
No, Julian. You've done plenty to damage your own reputation. My interest is simply to make the situation clear to the RISC OS public. You're right, I did slightly misquote you, but with almost the same meaning.
I've given many, many opportunities to back up things which you've said - but you've declined to do so. If you post misleading information, then it is not at all surprising that someone will correct you, or ask you for at least some kind of proof. It's only yourself you can blame for that.
I'm not sure how opinion comes into these things - for example, you claimed elsewhere that the Omega has a later Southbridge than the Iyonix - your opinion has little to do with this - in fact, you had only to open your machine to see if this was true or not.
And yes, I _do_ consider compilers a discussion of a technical nature. They are one of the most complex programs you can write. They are _extremely_ technical.
As for the issue which sparked this off - it's simply unacceptable for you to say something is "much faster" without any kind of proof _whatsoever_ - again, which I gave you plenty of opportunity to present. If you want to demonstrate some instance in which GCC 3.3.3 is much faster I would be most interested - however, nbench is an excellent representative of typical program behaviour, and it was run under the same conditions on the same machine, so it is extremely strong evidence that it is not.
Finally, I will question one more of your claims - that optimised StrongARM binaries will be faster on an XScale than ARM6 ones - is this true? I don't know. Again, I ask you to show some references for this, or evidence. I fear that once again, you will not bother, and we will once again be questioning your stance on telling us what is and isn't accurate.