In a word, this is correct. The "non-responsive" category in the polls presumably contains not only undecided but also those who will abstain or cast blank ballots.

There's every indication though that abstention will be much lower than in 02, which means by default that there is a high degree of indicision, and this is not uncharacteristic for a multi-candidate field (and especially for a 1st round presidential election. Exit polls in 02 reported >20% made their decision in final 48 hours).

And there are those who might change their minds.

Carroll of CSA notes in an interview in today's Monde that he still finds a very high degree of volatility in Bayrou's supporters -- up to 50% might still change their minds. That could mean a late surge for LePen (if, as the UMP apparently thinks, those soft Bayrouistes are protest voters from La France exasperee), it might mean a late surge for Sarkozy (as Carroll seems to imply, with no real basis), it might mean a late surge for Royal (which I don't expect it seems better supported by Carrol's own CSA polling of the last few days than the other hypotheses.)