The missing piece of the puzzle was how to pin out the connectors. At the time, I didn't find a lot of information on existing standards, other than a mention of one system that allowed use of a straight through cable to make pin-for-pin connections, and a rollover cable for null modem connections. Thinking on this, I decided it was not for me, as it would require multiple ground connections, Y connections, or some assumptions about the characteristics of the equipment to be connected.
I therefore decided to roll my own. Since ordinary 6P4C modular phone plugs will mate with the 8P8C jacks (albeit mashing on the wires a little much), I wanted to allow use of phone cables as patch cables as well. Some such phone cables only contain two wires (the center pair). Therefore, I elected to arrange the signals so even such wires would yield a useful (if minimal) link. So I made ground and TX the center two. Another wiring of the adaptors makes them ground and RX, for a one-way null-modem cable. The next two out got RX and DTR, so a common 4-wire phone cable can provide two-way connectivity, along with a signal lead. Working outward, I continued to add signals, in what I considered a logical order of precedence.
I shared this wiring with several friends, including rmd, who published it on her site as well. I figured the more people who used my scheme, the more convenient it would be. The document is still available on my web site.
Fast forward to today. I was looking up pinouts for Cisco console connectors, thinking I could just make an adaptor for that (Cisco uses 8P8C connectors for the console connection on many of their products, and uses them with rollover cables and adaptors, or all-in-one cables to connect to DE-9 serial ports).
I found that, over the years, a few different standards for this sort of connection had emerged. Cisco's scheme, documented here by a Zonker that is apparently not rmd, uses two ground leads in the center, and a fairly symmetrical pinout around it. Another scheme, also based on rollover cables, is the Yost one, which became fairly popular when it was published in a Unix System Administration Handbook. Dell had their own bizarre scheme for RS-232 over 6P4C.
Today, there's actually an international standard for this sort of wiring. It's EIA/TIA-561, often referred to as "RS-232D". How do you think it is wired? It doesn't use up multiple wires for ground and all signals are straight through. In fact, it looks very like my old scheme, although the pinouts don't match exactly. I assume the thinking behind it is similar. I am rather pleased with the indirect vindication. However, I'm left with the decision to just keep my existing adaptors the way they are, or switch them all around to match the new standard (and place a note on my web site explaining the change).