A couple more detailed comments/questions: 1. While the way you cast seemingly comforting terms like 'reuse' in new settings is a nice linguistic hack, it is a bit double-edged. For example, I gradually grew aware that I think of 'reuse' as the assumption that people can't see inside encapsulated layers like classes. That would be like a C++

.h

file that doesn't include implementations, or a Java

.class

file distributed in binary form with just APIdoc-generated comments. Whereas you're assuming that you can see inside classes, you just can't (or choose not to) modify them directly. Which leads naturally to the DOM model of hanging listeners and overrides on the skeleton of a given set of classes. I wonder if this distinction is worth explicitly drawing out. Otherwise you're bringing in distracting associations for your reader. (I've had this issue with past papers as well, now that I think about it.) This sort of reasoning is why I lately try to avoid the terms 'reuse' and 'abstraction' (and a few others). The statement of the principle at the start didn't immediately paint a crisp picture in my mind. It wasn't obvious why "fire up your text editor and modify a few lines of code" was not following the principle. (https://xkcd.com/722 came to mind.) At the same time, I could imagine other readers with a different background seeing it and thinking, "this is why we have late binding/dynamic dispatch/polymorphism/inheritance/..." So depending on the reader, the principle as stated will seem either trivially satisfied or a nice ivory-tower goal but not very actionable. By the time you return to it in section 5, my sense of your meaning is a lot more rich. But perhaps you should start with a problem rather than the principle? Hmm, which arguably is what you're doing in sections 2-4, but I didn't realize that at the time. Maybe you need to condense the problem down a bit? Speaking of that, I found Table 2 quite confusing. Coming from a Lisp background, I found myself wondering if you were just looking for some maximally-flexible way to specify lexical and dynamic scope. It was only when I got to section 3.3.2 that I realized that

β

may be deep inside the implementation of

α

. The colons obfuscate that because they make the pattern seem like a superficial use of an existing type. 2. It's hard to discuss a general term like 'reuse', particularly after bringing up economic considerations, without causing me to wonder precisely what is difficult about modifying some code directly rather than adapting or extending it. Might the paper benefit from a focus in a specific language's operational eco-system? For example, if you were speaking in terms of Python you'd be able to say something concrete like, "it's useful not to modify alpha directly because it comes from a package downloaded from pypi and you don't want to fork the upstream package and be responsible for merging in all future changes." Newspeak or Beta may have different reasons, and I'm not familiar with their eco-system. Why is it not an option to modify these classes directly? What exactly causes the network horizon in these situations? (I have my own answers to these questions, but I'm not sure if my answers match yours.) 3. A slightly less subtle way to beat my own drum:

...“available for use” meant that a module’s content should not be modifiable by its consumers, promoting uses such as caching, verification, etc.

Can we verify use even in the presence of modification, based on type checks, contracts, ahem tests, etc.?

The question of "what audience the paper was intended for" is also a fascinating open question. In fact, you're the first person I'm aware of who has come to this material "in the wild" and tried to make sense of it without any prior connection to the community behind it, so anything you say about problems in its presentation carries a huge weight. There's a huge "chicken and egg" problem with the ideas, which I think you're drawing out - you say "Perhaps you're writing for people who think reuse is a) great, b) attainable, and c) just a matter of sufficient foresight". But I think as you and I well know, no such person could ever possibly be convinced by material like this - and in fact would never read to the end of it. It's only because you had a substantial resonance with the ideas already and had independently constructed your own incarnations of several of them that you had the patience to wade through it. I think we have to say that in these cases we are writing for a form of "nonexistent ideal reader" - that is, what Lakatos would call the inhabitant of a "rational reconstruction of history". From the vantage point of the next century, when these ideas are universally accepted (!) some historian of science would try to reconstruct the intellectual trajectory that some wholly imaginary "central" member of the field took in order to navigate from the old ideas to the new. And this paper, I guess, is addressed to that imaginary member. And perhaps whilst there's no single person to whom the entire connected narrative is useful, the paper could be useful as a "grab bag" of arguments - that is, if one finds people in the wild who believe X, Y and not Z, one could then direct them to the bit of the paper where Z is established. And finally, the paper has to be useful to ourselves in terms of even articulating to ourselves what our arguments are. It's fantastically easy to get lost and forget how on earth it is that ones means connect to ones ends, if the chain of links is as long and as diffuse as it is. And after finally - the paper is useful as a "venue for conversations". One particularly marvellous result was that when I presented the paper, a couple of the the (about three or four) people in the audience who seemed to get the material started a conversation and came to realise that there were connections between their own research programmes that were indirectly suggested by it.