Of course, the Lib Dems would prefer another voting system, "Single Transferable Vote" (STV), but a referendum on Alternative Vote was the best that could be secured.
Alternative Vote is fairly similar to FPTP. The difference for voters is that instead of putting a single cross on the ballot paper, they get to rank the candidates by writing numbers "1", "2", "3"... That's called "preference voting". So, AV is FPTP with the additional feature of preference voting. In fact, in an AV election with fewer than three candidates, you're left with FPTP: after you've marked your first preference, your second is obvious.
Single Transferable Vote looks the same as Alternative Vote, except that constituencies are larger, and have several MPs in each; perhaps three to five. For example, Denis Mollison of Heriot-Watt University) has produced a map of proposed boundaries that you can see at http://www.ma.hw.ac.uk/~denis/stv4uk/ Just as AV reduces to FPTP in some circumstances, so STV reduces to AV when you're there's only one winner - in a single member constituency. In fact, a few of the constituencies in Mollison's proposal (like Shetland and Orkney Isles) only elect a single MP. In those constituencies, the voters are actually using Alternative Vote.
So, there's a clear sense in which AV can be considered as a partial victory for advocates of STV. One of the two features is won. At a later date, multi member constituencies could be added.
The question then, is should advocates of STV support AV. To me, the answer is a clear yes. If the referendum is lost, that won't create a clamour for STV, it'll simply let people argue that the electorate don't want preferential voting. On the other hand, after a couple of AV elections, people will see that preferential voting isn't hard, doesn't produce crazy results, and is actually much simpler for voters because they don't have to worry about tactical voting.
At that point, we'll be ready for a debate about multi-member constituencies. It won't be clouded by claims that preferential voting is complicated. It'll simply be a debate about the balance between proportionality of representation on the one hand, and the constituency link on the other. Mollison's proposal looks to resolve that quite nicely. It should produce much more proportional representation than FPTP, or even AV. On the other hand, the constituencies look quite natural, being based on traditional counties, and being small enough to retain a good constituency link.
AV isn't inherently proportional; it can be more -or less- proportional than FPTP depending on the distribution of voters. However, several commentators have suggested that the UK electorate are distributed in such a way that AV would, in fact, produce a more proportional result in general elections. And that means more coalition governments.
My guess is that only a coalition government is likely to produce further electoral reform. So a win in the AV referendum, means that future governments are more likely to be coalitions, and therefore more likely to give us STV. OK, I guess it's unlikely that the next government will introduce a new electoral system. That's probably not something that happens twice in a decade. We'll probably have more than one AV election. But, if we don't get AV, the next government is more likely to be a majority government with no interest in electoral reform, and we could be waiting another century for STV.
Finally, a quick look at history. Votes for women weren't achieved in a single step. Women over 30 were given the vote in 1918. The sky didn't fall in, and ten years later women got the same rights as men.