The problem wasn't so much 'interference' as CFL's intorerance to the "choppy" output of a Triac, and the random current trickles from a CFL that made it think you were toggling the local power switch on-off-on (which screwed up the relay-based "appliance" modules).
In theory, the "appliance" modules had a jumper you could cut to disable that. In reality, it seems like either the modules sold by x-10.com didn't implement it correctly.
From what I vaguely recall reading somewhere, the x-10 baseband IC had an embedded 555 & required a few capacitors & resistors to set the on-off-on trigger-timing to something like 500ms per state... but the modules sold by x-10.com either used out of spec components, or omitted them entirely, so even transient pulses lasting a millisecond or two could trigger them.
Assuming the "phantom on" problem WAS due entirely to omitted/wrong passive components, ones that implemented the external components properly (maybe tweaked to require a longer intermediate 'on' and or 'off' time to filter out trickle currents) could have fixed x10 (as a standard) to work fine for another 50 years.
Ditto, for Insteon. Insteon was a graceful evolution of X-10 & their light switches were (IMHO) *way* better, but Insteon's greed kept it from ever achieving mainstream success.
Personally, I wish "wi-fi" LED bulbs had X-10 and Insteon capabilities built in to fall back on. Granted, X10 was becoming painfully slow (back in the 1970s/80s, people didn't expect to have 20 individually-tweakable lights in a single family room), but Insteon's speed was fine.