The use of TEX and LaTeX for scientific publishing will only succeed if we adapt to the concept of `generalized mark-up'. Only in that way an author can fully concentrate on the writing of his manuscript, and leave the typographical intricacies to the publisher. Much has yet to be learned by everyone involved: authors, editors, and publishers. We can learn from each other: publishers can learn from authors what is wrong about their macro packages; authors can learn from publishers which rules of proper typesetting they violate.
It would be a good thing if the processes of consultation and education were formalized in working groups, or something similar. I know there are publishers who are reluctant about formalized consultation, because they are afraid it slows down their efforts to get going with TEX and LaTeX. However, it is my opinion that they can only benefit from participating in some form of formalized consultation. The introduction of TEX and LaTeX will take time anyway, and by listening to the authors, publishers may be prevented from releasing macro packages which are not accepted by the authors.
Authors could in such negotiations stress that they will only opt for some form of generalized mark-up, because working truly camera ready would take too much of their time. An inventory could be made of what authors have to learn about the proper mark-up of mathematical texts, and courses could be set up to educate the new generation of authors. I think that some time from now universities will provide introductory courses in LaTeX, just as they are now providing introductory courses in computer programming.
In the mean time, LaTeX users can enjoy this BibTeX style, which extends the concept of generalized mark-up to such a horrifying thing as the compilation of a list of references.