Great scientific discoveries are often named by the last names of scientists that made
them. For example, the most popular asymmetric cryptography system, RSA was
discovered by Rivest, Shamir and Adleman. Another notable example is the Knuth-
Morr\(is-Pr\)att algorithm, named by Knuth, Morris and Pratt.
Scientific papers reference earlier works a lot and it's not uncommon for one document
to use two different naming conventions: the short variation (e.g. KMP) using only
the first letters of authors last names and the long variation (e.g. Knu\(th-Mo\)rr\(is-Pr\)att)
using complete last names separated by hyphens.
We find mixing two conventions in one paper to be aesthetically unpleasing and would
like you to write a program that will transform long variations into short.

The first and only line of input will contain at most \(100\) characters, uppercase and lowercase letters of the English alphabet and hyphen (-, ASCII \(45\)). The first character will always be an uppercase letter. Hyphens will always be followed by an uppercase letter. All other characters will be lowercase letters.

The first and only line of output should contain the appropriate short variation.

riseoj 작성

출처 COCI 2009/2010 Contest 4