In addition to the modem control signals, USB adapters also really mess up the timing of the signals. Some RS-232 hardware does not use any flow control, and relies on character timeouts to detect end of message. USB ports packetize the data, and can break up the packets in unpredictable places, so that delays can appear in the message, causing the other side to detect an end of message too soon. Also, some equipment uses the control lines such as RTS/CTS, DSR/DTR, RI, and DCD for I/O purposes. In DOS, it wasn't that hard to get somewhat decent timing to create and/or detect signals of specific duration with these pins, using them as essentially free digital I/O. With a USB to serial adapter, especially under Windows, it is nearly impossible to get any precision timing. Add to that another layer of VM, and it just gets worse. How bad this is really depends on the chipset and drivers. I've had more luck with PL2303 based devices than FTDI along these lines, but for some applications, USB to serial just won't work.

Well, the meanings of words change, anyway, but not always for the best. At best, I would have called this a biomechanical cat, not a bionic one, but I'm in my 40's, so new uses of old terms bothers me more than some.

Bionic = Biological + Electronic. Where's the electronic part of all this?