Envoyer par SMS: System Modeling of Author’s Image in Texts of One Speech Genre