Hohpe argues that the useful architect is the one who rides the building's elevator: at home in the engine room, able to sit in the room where the budget is decided, and paid to translate between the two. Of this book I also read a commercial summary rather than the original, and it is worth saying why the damage is smaller here: the book is a set of short essays whose titles already carry the idea.
What I took from it
The elevator itself, which is the best description I know of what I do when a project goes sideways: it is rarely technical capability that is missing, it is someone who can turn a technical constraint into a business consequence, and back, in the same meeting.
The idea that architecture sells options. An architectural decision is not the one that solves today's problem but the one that keeps tomorrow's door open at a reasonable price, and that price can be named. It is the most honest way I have found to justify work that produces no new screen.
And a warning I repeat: systems nobody retires pile up, and an organisation that never switches anything off ends up living with things that no longer serve it and still have to be maintained. What gets turned off is part of the design, not an afterthought at the end of the project.
What I left out
The whole book, again. And here the loss is concrete: Hohpe writes from having been in those rooms, and the anecdotes are the material that convinces an executive. The summary keeps the conclusions without the stories, which is like keeping the punchline. It is on the list to read in full, and until then I will not quote its finer points.