re math, there is also what i call the "caucus effect." as pointed out by an ed schultz listener yesterday, caucus votes understate what the popular vote would be if extrapolated to the primary format. and, barack won most if not all of these, so should be "assigned" additional popular "votes." of course, the Her Royal Clinton folks can make the argument that their demographs didnt caucus. it's all a bunch of lawyer-parsing anyway. she hurt herself with the sd's by insisting on this popular vote crap, but the problem is that a lot of folks apparently bought in, and mccain is now continuing the ruse.

From Real Clear the actual popular vote numers are : 17,773,626 Obama and 17,822,145 Hillary. Hillary by 48,519.

Your numbers include estimates from caucuses but the astrict in that number should be reduced by 50,000 for Washington since they had a primary. The astrict reads as follows:
*(Iowa, Nevada, Washington & Maine Have Not Released Popular Vote Totals. RealClearPolitics has estimated the popular vote totals for Senator Obama and Clinton in these four states. RCP uses the WA Caucus results from February 9 in this estimate because the Caucuses on February 9 were the “official” contest recognized by the DNC to determine delegates to the Democratic convention. The estimate from these four Caucus states where there are not official popular vote numbers increases Senator Obama’s popular vote margin by 110,224. This number would be about 50,000 less if the Washington primary results from February 19th were used instead of the Washington Caucus results.)
http://www.realclearpolitics.com/epolls/2008/president/democratic_vote_count.html