This is about publishing in journals. I just had a (junior) coauthor suggest we do something I absolutely hate as a reviewer and editor: respond to a comment by a reviewer only in the response letter, not by changing the article. Here's my reply, which another (well-known) coauthor supported enthusiastically, so I thought I'd share:
I disagree that it's OK to just argue in the letter and not alter the paper, except in the extreme (though sadly not that unusual) case that the reviewer is just completely wrong on a point. For a good paper, editors and reviewers may tolerate people who argue without adjusting the paper, but no one likes it. It appears smug, and not properly appreciative of the effort reviewers go to, or the collaborative nature of publication. In general, even sometimes when the reviewer is completely wrong, it is a good idea to assume that other readers will make the same mistake, and to make some change, even if it's just an additional two-word phrase or citation, to make the mistake less likely the next time.