I recently picked up a RT6856T off eBay for a good price but it was pretty yellowed so I decided to try to retr0bright it. This would be my first time doing this so I looked at the original formulas and process as outlined on the original
retr0bright page but after a bit of Googling and too many YouTube videos I found multiple people just using professional hair developer cremes or liquids to accomplish the same thing as they are usually ~12% hydrogen peroxide.
The success of the retr0bright really seems to come down to a good % of hydrogen peroxide and UV / heat to promote the process.
The creme like the original retr0bright process would be applied directly to the surface and then covered by clear cling wrap but some people warned that unless you get it evenly coated the part or go out in and kinda squish it around during the process you may get blotchiness and uneven results. The clear developer is liquid and people seemed to have success just submersing the part in a tub with developer + water mix covering the tub in cling wrap and setting it out in the sun for awhile.
I decided to use the
Salon Care 40 Volume Clear Developer since the keyboard was fairly large and I didn't want to worry about blotchiness in the results. I used a storage bin that I had in my garage and dumped in the 32oz of the developer and added water. I weighted the parts down so they wouldn't float and set it out in the California sun.
While the keyboard plastics were being brightened I decided to give a good clean to the keys and the base keyboard since it was pretty dirty and grimy. I pulled all the keys and cleaned them in a dish soap + water solution in an ultrasonic cleaner. I also decided to coil the extremely long keyboard cable to help keep it manageable for the limited desk (or floor space) I have for the SGI at the moment.
Overall I am really happy with the results. I believe that I diluted the solution with too much water as after about 6 hours in the sun the plastics were still a bit yellowed, but still a big improvement, so I ended up setting them out in the sun for a second day and the results were much better. I believe with a less diluted solution a single day would have sufficed.

2xR12K-400mhz, 2gb RAM, MXE, Digital Video, Fiber Channel

2xR10k-225mhz, 2gb RAM, SI, SE, PCI-Cage w/ CAD Duo
NeXTcube (040, 25mhz, SCSI2SD v5.1, NU_IO, 16mb)
NeXTstation Turbo (33mhz, N4000A 17" Mono MegaPixel monitor, NeXT Non-ADB Keyboard/Mouse, 32mb, SCSI2SD v5.1)
NeXTstation Turbo Color (33mhz, NU_IO, NeXT ADB Keyboard/Mouse, 128mb, SCSI2SD v5.1)
Sun SPARCclassic (Dead PSU :|)