The importance of copyediting a scientific paper D. J. Bernstein 2005.05.04 ``Copyediting, proofing, reference checking, formatting of images, tables and data sets, and other functions performed by publishers add real value to manuscripts,'' says a typical publisher. Several weeks ago I send a paper to a publisher, Springer-Verlag. Today I receive a revised version of the paper from Springer. Let's see what value has been added to my manuscript by Springer's copyediting. One set of changes is predictable. I don't capitalize titles except for the first word. Springer insists on capitalizing almost all title words. Apparently big bold titles with only one capital are much more difficult to spot than big bold titles with two or three capitals. That's also why the title of the first section should be changed from ``Introduction'' to ``Introductory Comments.'' This adds real value to the manuscript. The first change that surprises me is at the top of the first page. I originally put the title onto one line: The Poly1305-AES message-authentication code Daniel J. Bernstein Springer decides to split the title before the hyphen: The Poly1305-AES Message -Authentication Code Daniel J. Bernstein Wow, that's really valuable for the reader. Thanks, Springer. I next notice that Springer has changed my theorems in Section 3--- Theorems 3.1, 3.2, and 3.3---to Theorems 1, 2, and 3. Gee, thanks, Springer. As a reader I _love_ it when I'm looking for Theorem 3.2 and find a theorem labelled as 2 instead of 3.2. Really adds confidence that I've found the right theorem. I love it even more when a paper says Theorem 2 and someone cites Theorem 2 and I have no idea which section to skim through. Really saves time. Springer also makes random changes to the spacing between words in several paragraphs. Here's one of my paragraphs: ----- There are several reasons that Poly1305-AES uses nonces. First, comparable protocols without nonces have security bounds that look like C(C+D)L/2^{106} rather than DL/2^{106}---here C is the number of messages authenticated by the sender, D is the number of forgery attempts, and L is the maximum message length---and thus cannot be used with confidence for large C. Second, nonces allow the invocation of AES to be carried out in parallel with most of the other operations in Poly1305-AES, reducing latency in many contexts. Third, most protocols have nonces anyway, for a variety of reasons: nonces are required for secure encryption, for example, and nonces allow trivial rejection of replayed messages. ----- Springer drastically reduces the spacing on the third line and shifts words from subsequent lines. The reduction is enough to make words seem to run together, but not enough to actually fit another word onto the third line, so the last word ends up poking into the right margin: ----- There are several reasons that Poly1305-AES uses nonces. First, comparable protocols without nonces have security bounds that look like C(C+D)L/2^{106} rather than DL/2^{106}---here C is the number of messages authenticated by the sender, D is the number of forgery attempts, and L is the maximum message length--- ----- Another highly valuable improvement for the reader. Thanks, Springer. Springer also uses slightly longer pages than it announced in its public ``llncs'' information. My original paper has twelve graphs split across two pages, followed by two carefully placed notes. The change in page length means that there is extra space below the first six graphs--- which I might not notice if it weren't for Springer _moving_ my notes _above_ all the graphs, bumping the first six graphs down into the extra space. Gee, thanks, Springer. You're really promoting the progress of science. I don't know what I'd do without you and your expert editing. Last, and quite clearly least, Springer changes commas to dashes inside citations. For example, at one point I write ``[31, Section 3]'' to cite Section 3 of the 31st paper in the bibliography. Springer changes this to ``[31--Section 3].'' Thanks, Springer. I'm amazed at how much value you've added. Is there any way I can arrange for universities to send you more money to fund this incredible service? To discover all these valuable changes, I have to spend a few hours carefully comparing my paper to Springer's paper, line by line. Springer never sends me a list. Obviously they don't want to spoil my fun. Once I'm done, I send email to Springer telling them how much I appreciate these changes. Thanks, Springer. Perhaps Springer will make other valuable changes to my manuscript---but I'll have to wait for the printed version before I can admire those changes. The first round of changes took several weeks, as you'd guess from the complexity of the changes; Springer wants to meet a particular deadline; so there's no time left for Springer to send me anything else. Too bad. I'm really looking forward to seeing how much more value Springer can add to my manuscript. But, even if there are no more changes, I'm so happy with the changes so far that I feel compelled to say this one more time: Thanks, Springer!