Guid compression

gistfile1.cs

static string compressGuid(Guid guid)
{
    // "MspzuC2oLkKjoUEgZrbJFg=="
    return Convert.ToBase64String(guid.ToByteArray())
        .Substring(0, 22)
        .Replace('+', '-')
        .Replace('/', '_');
}

static Guid expandGuid(string compressedGuid)
{
    string base64 = compressedGuid
        .Replace('_', '/')
        .Replace('-', '+') 
        + "==";
    var bytes = Convert.FromBase64String(base64);
    return new Guid(bytes);
}

Above method is called "compressGuid", but doesn't it convert a GUID, which usually takes up 8 bytes (plus object wrapping), into a base64 string, the example above being 24 chars long? (So either 24 or 48 bytes, plus object wrapping, depending on whether .NET is clever about storing strings with no Unicode in them)

Agree, @trejkaz , however I believe compressing Guid string representations could be useful for more space-effective tagging of transactions/requests in logs (where GUID woud be represented with it's ToString() result instead of 8 bytes in memory). Frankly, I don't think this can be done more efficiently, because it's already raw bytes represented with human-readable characters, because it's not Guid's ToString being converted, but instead it's ToByteArray().

Perhaps, smaz could do a better job. I never tried it though.
Or, use more of the alphabet! However, it definitely defeats the purpose, because the more of the alphabet is used, the less distinguishable the end-result becomes.