This is because everybody at McDonalds knows their job (in theory). The processes and roles are assigned and taught, so that you don't need the best chefs of Europe to cook or the finest MBAs to manage.
But who wants a software company that's like McDonalds? Most people, the argument goes, do not like variability. They want to know what they're getting, to know that when you promise something you'll deliver it. Every time. Then they'll keep coming back with their requests and changes, and won't jump ship to someone else.
This doesn't mean that there's no room for creativity. In software there has to be. But it means that your business processes (requirements, estimation, coding, testing, billing, rollout) are all standardized and do not lower the quality of your work.
In some ways, I guess this reflects the idea that when someone hires you to write software, they're not just buying the final product they're buying the experience. My local entertainment powerhouse, SheepWorld, knows this. You don't just get to see a couple of sheep, a dog, and a human. You get an authentic Kiwi farmer with his laconic sense of humour and natural sense of theater. You get a show.
Similarly with software, a customer buys not just lines of code but the process of acquiring them. If they have a good time, are treated well, and feel like they got what they asked for, they'll come back. And if they get the same happy experience, they'll come back again.
My uncle's example was a hair cutting place he went to. The first time he got a 40 year old guy who talked and joked as he cut, and my uncle had a great time. The next time he got a 25 year old woman who was silent as she cut. He tried again, and got a 30 year old man who sang along to the radio as he cut. There was no fourth time.
Personally, I'm trying to get on top of editing and come up with a process that makes it franchisable. In theory this is the publishing company's job, but O'Reilly has been very organic about the whole process. Editors have almost complete free reign in how they go from idea to final manuscript, and so there are no training manuals for new editors.
I've been picking it up as I go, but my goal is to be able to turn the entire process into a pipeline. Every step (feasibility, contract, development, tech reviews, marketing reports) should have a standard process I can follow. I lose time and hair, not to mention lower the quality of my authors' experience and possibly the books, by having to reinvent everything anew each time.
I can't tell you how enormously stressful it is to be at the bottom of this huge learning curve. Thanks to the recession and the loss of several coworkers, I now think of books in terms of jobs. "If this one is a stinker, how many people's salaries have I just wasted?" And every time something a book slips its deadline, I wonder about the costs of slippage--book buyers from the major outlets have fixed budgets and buy for a certain window, and if a book slips they cancel their order and it can be months before they can reorder from another month's budget. Did I just lose someone their job because I didn't get the illustrations to the graphic designer in time?
Getting editing to the point where it can be franchised means that I'll stop making my authors miserable, I'll stop making ME miserable, and I'll hopefully be able to get more and better books out. A nice dream, now all I need is a lifetime to make it happen