My friend and colleague Peter Adams just posted a nice essay on the early history of computing in our office. Some day I'll have to write my own version with the more recent history! Peter mentions the early use of the UNIX OS (troff was used for many years for typesetting the journals). This was originally run on Digital VAX hardware, then a series of Sequent machines, then (when I started) we began to replace those servers with Suns, and more recently everything's migrated to Linux. But still essentially UNIX at heart. A lot of what our editors do is still on the command line, and the referee and manuscript data in particular (going all the way back to 1974, plus a small number back-filled from before then to handle later errata) is pretty much completely accessible from a variety of command-line-oriented programs and scripts.
Of course, that command-line culture has been difficult to convey to new editorial and operations staff when they come in, and some of it was really quite cumbersome - especially because until just a few years back it still required paper copies of manuscript information to circulate to the responsible staff member. I remember seeing Peter coming and going from the office carrying a large mail crate filled with manuscripts, to do some of his work at home. So we've made pretty much everything also available through web applications in recent years with links to electronic copies of the manuscripts and correspondence, and a queuing system to assign things around. There are still a few pieces left to put together on that; it's been a major project that I've been centrally involved in for quite some time now. Peter Adams has been a leader in that as well, demonstrating some web-based approaches to doing things before anybody else was quite ready for it. But we were able to get rid of paper copies of manuscripts and correspondence, and it's hard to even remember now what a burden that was.
Thanks, Peter, for the memories of the way things used to be here!