Fairly high on my list of "I want those hours of my life back" is time spent wrangling my LaTeX to compile correctly with the style files provided by journal publishers, and subsequently dealing with copyeditors who introduce errors while making my papers conform to "house style".
Back when the primary medium of a journal was a paper publication, I suppose maybe it made some sense for all the papers in one journal to look the same. But nowadays I acquire nearly all math papers electronically, and I assume the same is true for many if not most mathematicians. Is there any good reason for journals to continue to insist on a house style?
To be clear, I certainly understand that some minimal requirements are necessary. For instance, a journal that has a print version will certainly want consistency in font size and margins, to ensure a fair comparison of the lengths of different papers. And I recognize that copyeditors serve a useful function in general. But I don't see why a journal needs to require that I use their custom
.cls file that is (for instance) incompatible with standard
amsthm, numbers equations as (1) (2) (3) rather than the more useful (1.3) (3.2) (5.3), or requires enumerated lists to be labeled 1.2.3 instead of (i)(ii)(iii).
(I recognize that this question is not intrinsically specific to mathematics. However, it seems especially relevant to mathematicians, and the people who I expect may be able to answer it, such as editors of math journals, seem likely to be found here in greater numbers than on any other stackexchange site.)