I don't know whether it's better to implement socialism — in the sense of a transition from capitalism (or feudalism) to communism — using a "top down" or a "bottom up" approach. But to simply declare that a "top down" approach isn't real socialism seems like substituting ideology for pragmatism and a No True Scotsman fallacy.
It's acceptable to take an ideological approach towards a goal; it's much less acceptable to take an ideological approach towards the means to achieve that goal. Ideology is not completely inappropriate towards means: you obviously want to avoid any means that will compromise your goal, and there are some means we consider ideologically unacceptable on their own merits, regardless of their effects. Even if I were confident that we could achieve communism — a classless, stateless society free of exploitation and oppression — by killing millions of people with the "wrong" sort of psychology, I would consider such means ideologically unacceptable.
But the ideological limitations on means apply mostly to extreme cases, and it's usually possible to transform an "essentialist" argument against means to a pragmatic argument. It's easy to argue that we cannot be confident that killing millions of people with the "wrong" sort of psychology would indeed lead to a communist society.
It is, of course, entirely legitimate to debate which of a "top down" or "bottom up" means is better to begin the implementation of communism. But it seems intellectually lazy to simply declare one or the other definitionally wrong.
Communist: Socialism is great!
Skeptic: But the USSR and the PRC were socialist, and they weren't that great.
Communist: The USSR and PRC weren't socialist.
Skeptic: How do you figure?
Communist: They weren't great. Socialism is great by definition.
Few people (mostly Christians) make their logical fallacies so blatant. Usually one picks some definitional component to exclude the undesirable cases, but the choice of component is driven only to exclude the cases, and only because they are undesirable. One could just as easily say that the USSR and PRC were not socialist because they didn't use the English language; since we in the United States would use the English language, we would be real socialists, and socialism — real socialism — is great.
The point is that making a definitional case avoids making the causal case, or makes the causal case superficial. If we simply deny that the USSR and PRC were socialist, then communists need not examine their successes or failures in any depth or rigor, any more than we need to examine Hitler's successes or failures as failures of socialism.
The definitional case is not entirely inapplicable: it's not a No True Scotsman fallacy to argue that Apu Nahasapeemapetilon is not a True Scotsman, because Apu is not being excluded by virtue of liking sugar in his porridge. Likewise, it's not a No True Scotsman fallacy to note that "National Socialism" (i.e. Nazism) is not socialist: not because Hitler was an evil bastard but because his economic and political philosophy had no connection whatsoever with Marxian philosophy; indeed Nazi Germany was violently opposed to anything that had any connection to Marx.
But even so, the definitional case has to be made very carefully, very rigorously, and one must bend way over backwards to avoid bias suggesting the No True Scotsman fallacy. Even a small mistake, a small hole, makes the argument look like bullshit to an honest skeptic. Addressing the stronger rebuttal is always more persuasive.
It's better I think to see the USSR and PRC as failed socialism, rather than not socialist, as attempts that failed — at some point — to move society towards communism. If we find that they failed because of decisions made in the first five minutes, we still have to make a causal case for that finding (and conclude they were monumentally inept socialists), not a definitional case.
I object to calling the USSR and PRC "not socialist" not because I have an enormous admiration for those societies and believe they must be emulated, but because I think we have much to learn from them, both good and bad.