One issue that you may not be aware of is that standard USB provides 5 V, whereas RS-232 is spec'd for a 12V swing.
Wrong. USB-to-RS232 converters usually contain voltage converters providing (high impedance) +/- 12 V, typically a variant of the MAX232 chip. And even if they did not, RS232 is specified with signal voltage levels of +/- 3 V to +/- 15 V, so + 5 V on a signal line are perfectly ok. But 0 V, as provided by too simple adapters, is not allowed, it is simply undefined. Most PCs and some other hardware accept 0 V as "negative because not significantly positive voltage", but I would not rely on that. Another wrong assumption is that RS232 lines could deliver significant ammounts of energy. The COM ports in the original IBM PC and almost all clones could deliver several mA, sufficent for a mouse, but that's not in the RS232 standard. Still, all USB-to-RS232 Adapters I have seen supply enough energy to light the LEDs in a RS232 tester dongle.
Today I will gladly share my knowledge and experience, for there are no sweeter words than "I told you so". ;-)