The following test shows that the canonical names for some character encodings have changed from 1.3.1. In the case of UTF-16 that's probably just a simple typo. In the cases of UTF-16BE/LE it may indicate a more severe incompatibility in that the 1.4 converters produce/require a byte order mark where the 1.3.1 converters didn't.

import java.io.InputStreamReader;
import java.io.IOException;

public class CheckCanonicalNames {

    public static void main (String[] args) throws IOException {
    
        checkCanonicalName("ASCII");
        checkCanonicalName("Cp1252");
        checkCanonicalName("ISO8859_1");
        checkCanonicalName("UnicodeBig");
        checkCanonicalName("UnicodeBigUnmarked");
        checkCanonicalName("UnicodeLittle");
        checkCanonicalName("UnicodeLittleUnmarked");
        checkCanonicalName("UTF8");
        checkCanonicalName("UTF-16");
        
        checkMappedName("US-ASCII", "ASCII");
        checkMappedName("ISO-8859-1", "ISO8859_1");
        checkMappedName("UTF-8", "UTF8");
        checkMappedName("UTF-16BE", "UnicodeBigUnmarked");
        checkMappedName("UTF-16LE", "UnicodeLittleUnmarked");
    }
    
    private static void checkCanonicalName(String name) throws IOException {
        checkMappedName(name, name);
    }
    
    private static void checkMappedName(String alias, String canonical) throws IOException {
        InputStreamReader reader = new InputStreamReader(System.in, alias);
        if (!reader.getEncoding().equals(canonical)) {
            System.out.println("Error: name mismatch for " + alias
                    + " - expected " + canonical
                    + ", got " + reader.getEncoding());
            }
    }
}

No output on 1.3.1.

Output on 1.4 B83:
Error: name mismatch for UTF-16 - expected UTF-16, got UTF16
Error: name mismatch for UTF-16BE - expected UnicodeBigUnmarked, got UnicodeBig
Error: name mismatch for UTF-16LE - expected UnicodeLittleUnmarked, got UnicodeLittle