One of the reasons is that UML attempts to become a programming language.
By aiming to be able to generate full code actually UML tries to be a programming language. In my mind there is a big problem with a general purpose graphical programming language. In human history the written form of all languages evolved from graphical to textual (see also Visual Designs Don’t Scale). Alphabets proved to be more versatile and more expressive than pictures in capturing ideas. Try to describe any simple process in images. The funny thing is you still have to annotate the images with words. And the full text version with no pictures still gives you more details. Pictures prove to be good at sharing ideas and allowing people to visualize concepts. But in the end words are better at describing the fine details.
Federico Cargnelutti 
3 responses to “13 reasons why UML failed”
I disagree that languages evolved from images to text. Frankly, I don’t believe text to be a superior form of expression. Just look at all the scholarly debate surrounding bible interpretations, etc.
Then again, can we separate images from words meaningfully? For instance, you read the word cow and what happens? You have the image of a cow popping up in your head. What is a cow in the absence of this image? Does any such thing exist?
What is a cow? Protons, neutrons, electrons? Subatomic particles? Nah. Those are empty abstractions. They tell me nothing about a cow. A cow is the thing standing in the field with the hay in it’s mouth, not the sum of some constituent parts dreamed up by some uber intellectuals some time ago. It’s the whole entire thing.
And when we think about the whole entire thing, what do we see? We see, well, the whole entire thing! All at once. Boom! It’s in your face. I am a cow! You don’t just read a cow from left to right and top to bottom, you read all of the information in a glance. Now that’s effective communication!
From what little I’ve read on the interpretation of Egyptian hieroglyphics, it seems this form of communication works in a similar fashion. The glyphs can supposedly be read in relation to all neighboring glyphs, not just those to the left and right (I say supposedly because I’m no Egyptologist!) I think this style of “reading” would call for an interesting combination of left and right brain activity and may even be superior, but who can say? We’ve moved away from it in any case. Not evolved, just moved away.
In summary, just because information isn’t readily accessible to the left brain doesn’t mean it’s not valuable information. In many cases the old adage is true: a picture is worth a thousand words.
Indeed, but can you make a cow based on what you see all at once?
Both, images and text, work hand-id-hand, one complements the other.