Would you rather ATI have the best performing single card(at any price, definitely not something cheap), but lose at every single price point below that like a company who will not be named did the last generation? Or would you rather them be competitive with the fastest opposition card for much cheaper while having something cheaper and or faster at every other price level?
Who are you supposed to be asking? If it is a hypothetical ATi fan, then using the last generation isn't a good one if you want the company to continue to exist. They were utterly destroyed in the marketplace when looking at the 4xxx parts versus nV's lineup, by a 2:1 ratio. While the ATi fans may have loved what they did, doing the exact same thing would be terrible business and help them move towards collapse. I'm not saying that is what is going to happen, but cheering on a strategy that is leading you to your demise just doesn't make any sense.
If you are asking any sane rational person they wouldn't really care if it was nV or ATi that offered the best choice for them. One of the big problems that ATi seemed to have last generation is the perception of value. Despite them clearly leading in price/performance when looking at raw benchmark data, they failed to be competitive in the marketplace in any fashion whatsoever. I'm not sure which element was the biggest factor, nV pushing PhysX, CUDA or perhaps the normally vastly superior nV bundles(likely some combination of all of the above) but whatever it was ATi needs to address it. I have stated before and I stand by the sentiment that it doesn't matter if consumers want a bumper sticker with their video card, if that's what they want, you give them their stupid bumper sticker to get the sale.
If the 5870 launched at $299 I think you would find most people would be overjoyed with it, and it wouldn't surprise me(that's actually a bit of an understatement) if ATi doesn't already have all the pricing promos ready at that level for as soon as nV launches. Right now the are supply limited, they could have launched at $500 and would have sold out(and that would have been the smartest thing to do short term, but it would have left a bad impression on some people). As it stands right now the 5870 ranks somewhere around fifth best solution in its' price range(4890CF, 275SLI, 4870x2, 4870CF) looking at a price/performance metric. Please try and keep in mind that the very same ATi loyalists saying those aren't valid options I can pull up examples of them touting the superiority of the 4870x2 over the GTX285 and the 4850x2 over the GTX280.
In terms of superior feature support, again, this is coming off as rather comical. A lot of people in this thread are talking about SSAA and angle independant AF, both of which are features the competition has had for years(in other words, it's only a valid upgrade point if you don't consider the green team). The major features of this part are Eyefinity and DX11 support, neither of which I will try to marginalize in any way, but DX adoption rates have always been very slow and it is
always going to be different, this time. If someone wants to buy a 5870 for DX11 support I think it is a perfectly valid reason, but in all honesty we do have far more PhysX titles on the market right now then DX11 offerings, and that will likely remain accurate for at least a couple quarters if not longer. Again, if people are looking to future proof then I can completely understand that, but this forum has been filled with the red team fans saying that you shouldn't ever future proof, buy for the games you are playing today etc, etc.
Pretty much, the 5870 is very much like a normal nV launch in many regards. It brings the top single GPU performance to the table, does so at a price point that does not make it particularly attractive when compared to the competition but it offers some features that have some interesting benefits in both the short and long term. I do not think the 5870 is a bad part in any stretch of the word, it just seems very much like what I would expect from the green team. Taking early profit margins(smart business), pushing features over performance and relying on technical superiority to sell its products in spite of its inferior price/performance compared to the competition. None of these things I consider bad, just not typical ATi(which given how their business model has been working versus nV, that may be a very, very good thing for them).