I figured I should wade in and say a few words, just because I find it hard to stay out of flame wars. Woohoo!
i) I talked to Andrew about the 4/10 score when the review was written, and after a long discussion he said he definitely wanted to stick by it. It wasn't an off-hand thought or a typo.
ii) I totally think Fedora is a desktop distribution. As Andrew says, it goes into RHEL for use as RHEL WS and RHEL Desktop, both of which are not aimed at servers. Please keep in mind that Fedora was called Red Hat Linux only a few years ago, and sold in bookstores the world over.
iii) GCC 4 does not, as yet, bring about any substantial speed improvements to the best of my knowledge. This is supported not only by our extensive tests (see the feature in LXF66), but also by others (see http://www.coyotegulch.com/reviews/gcc4/
for the quote "Is GCC 4.0 better than its predecessors? In terms of raw numbers, the answer is a definite "no".") That said - and this is important - I still think GCC 4 is a great release. It has built the foundation for a number of further optimisations that will come in the future, so I expect FC5 to really be able to make the most of it.
iv) The jumps from MySQL 3 to 4 and from PHP 4 to 5 are most welcome, but these are just version bumps. I am honestly sick of distros that do little beyond updating their packages, and version number superiority just doesn't impress me any more. Of course I'm glad to have more people using PHP 5, but to be honest it's way, way overdue - PHP 5.0 was released on the same day as FC3 Test 1, yet (presumably to avoid breakage in RHEL 4) didn't find its way into that release. I don't think Red Hat deserves any kudos for leaving PHP 5 so long...
v) Obviously it's not possible for Andrew to accurately predict the release of FC5!
Regarding the overall scores, I would still have scored Fedora higher if I were reviewing it myself. I too am thoroughly annoyed at the lack of configuration tools - and indeed the lack of any real innovation - but I would still probably have scored it 5/10. As I wrote in my column a few issues ago, we re-adjusted all our scores to really mean what they say - 5/10 is "average", meaning that the majority of distros out there should get that score. I consider FC4 very average as a desktop distro. On the flip side, I would never, ever use it as a server distro either, simply because it doesn't provide a good enough security infrastructure for my requirements.
Nobber: it's not the case that Andrew doesn't care about Fedora. If that were true, he wouldn't run it on so many machines. Instead, I think the negative review is more of a reflection on the general disappointment of the release - it's not what it could (should) have been, particularly given the long lead time before RHEL5 is frozen.
So, to sum up: I think 4/10 is low, but Andrew's the expert. Like Andrew I've been using Fedora for a number of years, and have been using FC4 since before Test 1 - but I would still have given it 5/10 rather than 4/10. The points you've raised are valid, and I'd like to think that Andrew will take onboard all you've said and let that give him a fresh perspective when reviewing products in the future. He has reviewed Fedora for us for several releases now, and has a Fedora book out with his name on, so there can really be no question of bias - what is he biased in favour of, exactly?!