The Apple Newton was a stupendous failure as well, as far as I can tell, only 300,000 were sold or roughly 60,000 per year. Palm basically shipped the PDA that everyone wanted by focusing on simplicity, functionality, and accuracy. Let's just add that to fix up the snarky selective quoting of history that papers over Apple's failures while emphasizing others.
I'd say there's a path dependency issue at work here. Previously tablets failed because both the technology wasn't there yet, in terms of weight, performance per watt, usefulness to justify their use over a small laptop. Before the Web taking off, if the stars had aligned in terms of technology availability (capacitive screens, powerful ARM chips, tiny radio chips, etc) Microsoft would have had a chance at launching something like an iPad, because of the lock-in effect of the MS ecosystem, but the Web really devalued the need to run Windows. The original iPhone took advantage of this, simply including a great, usable web-browser, Mobile Safari (without native apps) was already a huge value -- from day one, they had a vast content repository. Whereas companies like Palm had to launch with only builtin apps, and build up from there. The Web provided a bridge while Apple built up their native app ecosystem -- even without native apps, the original iPhone was quite for usable because of this.
But the iPhone provided years of native app content creation prior to the arrival of the iPad, and so out of the gates, it had two things going for it, the Web, and a huge iOS software catalog that is mostly exclusive to Apple. Now other people wishing to enter the Tablet space face the same proposition people wanting to enter the Desktop PC space faced: They can make a "from scratch" platform and face the issue of no launch software while their competitors have lots. iOS has become the new Windows monopoly in that regard. Yes, some people want iOS because "it's better", but a lot of people who've never touched an iDevice will want iOS because of brand and because of the huge software library.
So it is hard to evaluate Surface on the grounds of purely technology or design. I think Microsoft could produce the most advanced tablet ever with the best and most usable OS ever, and it would fail, because at this point, improvements in specs or OS design are of marginal value compared to third party software.
The mobile space has started to crystalize, it's become mainstream, people have become used to the form factor and operation semantics, and the players who were out of the gates first, Apple and Google/Samsung, have now reaped the benefits of having an ecosystem and widespread brand association. It will be hard for any third platform player to change this, even with enormous sums of money. (Ok, it doesn't help that Microsoft faces the innovators dilemma and continually tries to 'link' Windows to their mobile stuff, when consumers don't want Windows/Office. They want chatting, music, watching video, playing games, and taking pictures. MS is marketing their stuff like enterprise users should buy it)
Look at how long it took for Windows to be weakened, it took a combination of the Web and Mobile, huge disruptive movements, and Windows is still selling billions per year, so they aren't going to go away anytime soon, but they can't win by making a third platform with the same form factor/usage characteristics. iOS and Android are going to own that for another decade. At this point, they need to focus on something new, maybe it's XBox One/Kinect and the TV market. Maybe it's wearables. They need a new platform.