Congrats on being an early reader, b/c I spotted this on a fresh read Monday morning and zapped it! [I'm a IEEE Spectrum editor] This is a problem that's got me the odd time before—when you've read an article so many times, and something somehow gets duped in the transfer to the online content management system, it's easy not to notice because you think you just lost focus for a second and your eye jumped back to re-reading the same paragraph, not that it's actually the same para twice. You get as many eyes as possible on articles to prevent this from happening, but every now and then something slips through, especially when you have a very atypical layout and structure compared to our regular articles, because focusing on those things is stealing cycles. So that one's on me!
Ugh, should be fixed. Another transfer problem that always has to be catched manually. I've been putting articles online since the mid-1990s, and in 30 fecking years of watching ever more elaborate text processing, DTP, web and content systems evolve from hand-assembled HTML 2 through homebrew content-management systems to CSS/HTML 5 and the fancy commercial CMSs of today, and desktop processor speeds go from 66 MHz to 3 GHz, we still can't reliably guarantee things like italics or superscripts will automatically go all the way through from word processor to browser.
I honestly thought that problem would be solved way before e.g. automated interview transcription...
