Thanks for your comprehensive reply -- I hope you still have some hair left on your head (lol).
I experienced quite a few problems in getting this display to work. The displays I have are distributed by Vishay, but they are nothing more than Winstar with a Vishay stick on label applied to the flex circuit. For reference, these carry the p/n OLED-016002B-LPPN500000, and have a QC test date 04/05/17 followed by 145103 (possibly a production lot number?). The other flex circuit that attaches the glass display to the circuit board is labeled EO1602B REV.0 (I think this says it all). There's also some printing on the circuit board, but being under this flex cable makes it unreadable.
I initially experienced considerable difficulty getting any response from these displays as Vishay had them incorrectly spec'd as having the 0010 controller in them rather than the 0012. Once knowing this I eventually happened upon an initialization sequence that worked and was able to communicate via 4-bit parallel. My initialization file includes the 4-bit mode command as I wrote an all-purpose include file, and this command doesn't adversely interfere with SPI mode, so I left it in.
After success with the 4-bit interface, I moved on to I2C as this was the configuration I was planning to use. After spending several hours unable to achieve any success (couldn't even get an ACK response from it), I contacted Vishay tech support, who in turn contacted Winstar. I eventually received a reply indicating I2C isn't implemented in this display. I wasn't told why -- whether the problem is chip level or circuit board level -- but I can confirm it no workie. Nevertheless, if your display has a more recent production date, I'd be interested in knowing if your I2C interface functions. This does require shifting around some 0 ohm shorting jumpers, a process I would not recommend to anyone without considerable surface mount soldering skill.
Regarding the reset problem, I can confirm similar behavior on my Vishay displays. This occurs regularly when I load revised code into my target uC, and execute a soft reset. Something hangs in the 0012's registers, and it either doesn't re-initialize, or it diverts off into some erratic behavior. The only remedy is a hard reset via power cycle.
I haven't deployed this display in anything beyond a prototype, and in my application the device always starts from a power off state, so the display benefits from regular hard resets. Unless a newer version of the display was to become available with I2C and predictable resetting, I'm not planning to pursue use of this display (although I had considered using it in SPI mode and including a FET control circuit to cycle power to it).
Overall, I'm not especially impressed with the Winstar/Vishay OLED displays of any variety. I've used the 4x20 0010-based display in a radio controlled (WWVB) clock application, and have found its pixels are becoming irregularly illuminated, although I haven't experienced any abnormalities with the 0010 controller.
I've moved on to Newhaven OLED displays that are US2066 controller based. They're more expensive, but the I2C interface works and the displays operate at 3.3 to 5.0 VDC. Newhaven's NHD-0216MW display is about the same physical size as this Winstar 2x16 display, although I've not yet tested it. The other Newhaven OLED displays I've tested are top quality and the best I've tried.
I've also been evaluating some of the Winstar/Vishay OLED graphic displays (100x16 and 100x32), and without going into tedious detail can tell you that the 100x32 display's 2 controllers don't cooperate properly. The top half of the display works OK, but the bottom half corrupts the top half. It also suffers from resetting problems, so I've pretty much bagged the Vishay OLEDs and moved on to Newhaven.
Hope some of this helps.