Ken, I read a description of this technique by Ralph Lambrecht somewhere, which said to flex the CD when looking at the reflection. It would separate the different colors into fringes (similar to a rainbow) and would be easier to identify the unwanted colors. I tried it and it worked just like he described it. Just thought I'd mention it.
Yes, that original suggestion has certainly made its way around.
Note also that the CD/DVD test can be performed on a low pressure sodium sodium vapor safelight, such as the now discontinued Thomas Duplex, or the still available (as far as I know) Sebastian Darkroom Products
Model OC-012 safelight.
In every case you will also find faint unwanted bands of blues and greens buried in the otherwise overwhelming brilliant orange of the twin sodium doublet emission lines. Again, too faint to detect with the unaided eye, but still present.
In the LEDs the unwanted colors are the result of various dopants intentionally introduced during manufacture. In the LPS lamps the extra color come from the intentional addition of
Penning gases (neon and argon) to facilitate the initial striking of the lamp. These gases cause that weird reddish-purplish glow one sees at the beginning before the sodium has vaporized and begun emitting its characteristic orange.
The bottom line is that no light source is completely photo-safe. They all require some form of filtration to help stretch out the usable "safe" timeframe.
Rubylith is an excellent and inexpensive solution for red LEDs, as it completely blocks the blues and greens, but passes virtually all of the useful red. My filtered 635nm red LEDs have been tested safe beyond 60 minutes for Ilford MGIV.
Ken