Maintenance is not conflicting with change, it's a form of change and it lays foundation for more drastic changes.
I'm mystified by this. Why does maintenance have to be a 'conservative' thing? Probably 70% of software engineering work is maintenance of one kind of or another. Are those engineers conservatives? Do they 'resist change'? Unless you mean 'conversative' in a very restricted sense of that word.
> so well-thought-out, so vivid and so blatantly superior to the incumbent alternatives that they become inevitable. That is, society cannot help but enact them.
This is a version of 'technological determinism' (Google it if you need to). If there's any single takeaway from the last 50 years of scholarship on the history of technnology, it's that technological determism is false.
This kind of thinking is astonishing to me. "Technological determinism" is a historical term--i.e. reified in its era. The arguments ("stirrups enabled feudalism") are dated and simplistic and easily torn down. But to reduce a thought (discovery of technology influences human behavior, and can bring about its own creation) to a label ("technological determinism") and then dismiss it for historical reasons (older versions of that label were insufficient, therefore it is wrong), is sloppy thinking, imo.
>This is a version of 'technological determinism'
True, the growth of technology can't be predicted. There's no guarantee any particular idea will work. Many ideas will fail due to unforeseen problems, including in apparently unrelated areas like financing and social media and so on. Entrepreneurs may take up an idea and then change it beyond recognition (PG points out that start-ups usually change their product ideas). Each attempt to enact an idea is just that, an attempt.
Whereas in the past we built things before we understood why they worked (steam engines), or whether people would like them, more and more we'll be trying out stuff in theory (richly animated and exciting fiction) before trying in practice.
While this is true, it doesn't prevent a hell of a lot of our engineered world from just being retained while the innovations are piled on top. So there is a hell of a lot of close-to-static stuff that matters a lot -- probably more than the innovative works of any given day.
If the old layers were a bit dodgy, their abstractions will leak, foundations will shift and problems will appear in the layers above. And most things are a bit dodgy to start of with, and will not become a firm foundation until they have gone through incremental improvement and maintenance.
Yet in the long run we absolutely depend on change in order to adapt and survive. Therefore there has to be a rigorous way of reconciling these two principles. Without fudging ('maintenance often matters more...')
Maybe: visionaries will develop new ideas with no intention to enact them. Eventually a few will become so well-thought-out, so vivid and so blatantly superior to the incumbent alternatives that they become inevitable. That is, society cannot help but enact them.
This already seems to be happening in some areas:
e.g. moral improvements which come about via fiction, especially fantasy fiction
e.g. individual decision-making (it seems like we deliberate for a while and then actions take place automatically)
e.g. Project Hieroglyph (no idea how this is getting along but what a great idea)
http://hieroglyph.asu.edu/
...and of course all those brilliant videos on YT of engineering schemes for the future, e.g.
https://www.youtube.com/watch?v=LMbI6sk-62E