Was it because of this? I can see that causing some issues.
Perhaps it was complicated by aesthetic concerns: when you put these internals on the outside, there's some pressure to arrange them in a nice- looking way rather than in the most functional configuration.
Accessibility is improved, but lifespan is sacrificed. When the systems are on the outside they're subjected to much greater swings in environment (humidity, pressure, temperature, thermal expansion, and so on) that don't happen when they're inside the climate-control envelope of the building.
Don’t buildings like these usually have access panels for maintenance? Plus a lot of “fake” paneling, especially on ceilings, that’s easy to remove to access all the piping it hides?
But when these things are outside, now you need a crane instead of just ladders.
Even without access panels, opening and replacing some segments of drywall sounds pretty easy compared to working on a building’s facade.
Would that approach change the original design? Absolutely. But I'd see that as a positive. Not because I don't like the original design (I do) but because it would make a nice teaching example: if architects don't want their designs to be messed with, they'd better start making "how will it look and perform after x decades of use and weather" much more a factor in their decisions than they do now. I'm so tired of seeing new architecture that makes pristineness (while new) it's central design statement and has absolutely nothing to show for once that new car smell is gone.
https://www.architectural-review.com/buildings/pompidou-cann...
Does that all sound familiar to folks building backend systems? That is 'microservices' for architecture, a counter to 'monolithic' structures, using almost precisely the same motivations, and, running into the same exact problems: operational complexity.