UML is going down ? or is it the whole programming thing ?

At first I was looking at UML because I like it and would like to see more MDA and MDSD done. But I read 13 reasons for UML’s descent into darkness and the future doesn't seem to shine :


But since I am very skeptical about the conclusion that are generally drown from Google Trends, I searched a little more and it seems that whatever the subject "Java, SQL, HTML ..." the trends goes down more or less :

So, I still hope to see something good out of AndroMDA or Taylor.