You often have to get very creative about how to get things arranged in tight spaces.
So you are saying it's impossible for a rookie to either draw in 3D or note the dimensions of obstructions and make sure there is enough room for fitters to install? It seems like it would be fairly obvious that you would go for the most wide open spaces possible - do the best you can, and if it still looks like there isn't enough space then create space by decreasing pipe size or whatever. If you had memorized the standards and guide books you would know on hand what things you could do to change the variables and create more space. I'm not arguing with you to defend a point but rather to get you to elaborate.
I could write a computer program that could recall a terabyte of data on fire protection but it probably won't do as good of a design as Travis could.
If you accounted for all the variables I don't see why it wouldn't do a much better job (especially in terms of speed) than any sprinkler designer. For example, in the video game 'tron', you have a bike that has a tail - like a snake - that slithers around in a square and tries to box another bike-snake in -since crashing into a wall results in death-. There are computer algorithms that people have developed to search for the most wide open spaces, or calculate spaces between walls and plot a path, to survive. In the future most jobs will be automated, even jobs thought to require human creativity. This is because in actuality humans are still limited by their environment just like computers, we simply calculate for many times more variables.
Most of our advantages only come in handy for exceptionally fast continual learning of new variables. Computer programs don't have to adapt so quickly.
For example, to solve the tight spaces problem, you would write some code that calculated minimum distance acceptable from obstructions for it to consider drawing pipe there into your entire pipe calculation subroutine. You could also continually update the prices for various pipe types, their sizes, sprinkler types, their coverage area, etc. If you did it right, it would be able to 'trial and error' calculate various sprinkler systems in a building, then you would program a simulator that would test the effectiveness of the system. Leaving the program running for a while would no doubt arrive at the most optimal system possible for the variables you gave it in a quicker time than a human could, and even if not, it would definitely be for lower cost (just for the cost of the electricity to run the computer).
My point is that if a computer in the future can account for all the variables, there must be somewhere that humans can learn OF all the common variables (except super rare cases), since the primary goal of people publishing books on the subject would be to teach rookies these variables. What else would the writers of such guidebooks be doing? How hard would it be to put in a line of text that says "Find the most wide open space possible for pipe in tight spaces so your fitters don't have trouble."
The example given does seem like it might be left out of textbooks, it doesn't seem like it can't be incorporated into them with ease, it's not an idea that can't be expressed in writing, such as the 'proper feeling of weight to put on the brake of a car to slow it to a stop without causing unnecessary wear on the brake, screeching, or dangerous sliding such as on a slippery surface'. THAT is an example of something you can't teach in a book.
