EDIT: I don't own a HD5000 card, but the devs have stated here that the driver should be on par with the HD4000 driver for the most part. They share much of the code anyway.
And here we have the important part.
"the driver should be on par with the HD4000 driver for the most part."
So, for the most part it's on par with the HD4000 driver. Not quite, but mostly.
So assuming the open source HD4000 driver driving a HD4000 series card was sufficient, there's a chance that may be the case also with a HD5000 series card.
1/ What about when a HD4000 series card driven by the open driver isn't enough and is only sufficient with the closed one.
2/ What about when a HD4000 series card driven by the open driver is sufficient but not in a particular case where a HD5000 via the open driver isn't.
If I said, "buy this fantastic nVidia card. And it'll usually work!"
I'd expect to be asked if there's a better functioning alternative and if I was aware of it I'd mention it. Just like Michael did.
I'm not saying a 5000 series card driven by the open driver never works, but I'm comparing it to what you can expect from the best solutions out there. If someone asks what's the best open source friendly video card manufacturer I'm going to say ATI. If someone asks what the best option for a video card that has the least amount of odd issues I'm likely to suggest nVidia at this stage.
With any luck in the future ATI will make that position obsolete.
If I removing all subjective comments from article, I see a card which is working fine, running all the test without crashing, runing firefox... so the card is working!! but without video accel. Did michael mention problems during installation or crahes? bad desktop behaviour?
The bad numbers on 3d is because this card is not for that, equivalent maybe to a GT210 and not a GT220. I don't understand why is a need to kick fglrx so hard.
I don't understand why is a need to kick fglrx so hard.
I don't see fglrx being kicked so hard, just describing real world issues that it has.
If the objective is to point to the best behaving solution then I don't think that's fglrx at the moment. If the objective is to point to the most open source friendly product then that's obviously ATI but those who value that above all else are likely to be across the subject well enough themselves.
When fglrx is pointed to for having reliability issues those wanting ATI to be the best choice usually then want to turn the discussion away from it and onto the open source drivers as if they were completely sufficient on their own. I acknowledge that for some they may be though. When discussing which card is the best solution it shouldn't be about which manufacturer one wants to be "the bestest" but which cards are the least troublesome and most appropriate for a given scenario.
I have boxes with Intel, nVidia and ATI graphics hardware. My main machine is the ATI one so when fglrx stabilises I'll be quite pleased to say the least. My arguments aren't really which manufacturer I like best but it's which card will be the most trouble free.
If Michael wants to write an article about which is the most open source friendly manufacturer then we all know basically what it'll say. I don't think that was what the point of the article was, it's only one of the facets of the issue.