The popularly known Human Development Index (HDI) is obtained through linear averaging (LA) of indices in three dimensions - health, education and standard of living. LA method assumes perfect substitutability among the indices. We question its appropriateness and propose an alternative measure, which is the inverse of the Euclidian distance from the ideal. Following Zeleny (1974), we refer to this, as the Displaced Ideal (DI) method. Through an axiomatic characterization, the paper shows that the advantages in the DI method are the following. Uniform, as against skewed, development is rewarded. Through an ideal path, it signals a future course of action. These signify that a given increment in any one dimension, with other dimensions remaining constant, has a greater significance for the index at a lower level than at a higher level. In other words, stagnancy in the dimension that has a lower value is more serious than stagnancy in other dimensions. Finally, an empirical illustration has been done by taking the statistics in Human Development Report 2006. We strongly propose that the DI method be considered over the LA method in the construction of HDI.