oh
http://www.who.int/whosis/whostat2006DefinitionsAndMetadata.pdf
says it is an AVERAGE number of years expected
( remember an average did not the half way point - 50% do and 50% dont that is called a median = very common error )
average is balanced - if you have a dataset 1,2,3,4,1000 - th average is add em all up and divide by 5 which is around 202 and you wil notice four points are below the average and one above
The median - the bumber at which 50% is above and 50% below is 3
for the second part of your question - life expedctancy when they arent all dead such as those born in 1950 - try googing Kaplan Meier survival curves and censored data.....