Exactly. Many posts will go on about how having a good architecture will make it easier to maintain in the long run, and other such things, but it all comes down to one thing:

TIME

If your code is useful, it will be used, you will get revenue (hopefully!), and you will have the TIME to improve things, clean up code, write unit tests, and all of the other things that are good and proper in life, but which only indirectly benefit the end used. This is the fundamental opportunity cost for software.

Yet another example that any large, established, company will never knowingly introduce a new product that might damage the market for an existing product. That is why giving billions to one or two large companies to develop TECHNOLOGY X never seems to work. If you gave the same amount of money to companies with less than 50 people, you would have 12 different versions of TECHNOLOGY X within a year.

End rant.

All that I can picture is the classic 19th century drill tower with glowing magma spraying from the top, and lava-coated workmen running around cheering "It's a gusher!!"

Actually, in my mind, the workmen look a lot like Homer Simpson ...