There was already some text and schematics for Duet fan connection here :
http://otocoup.com/aql/#hlp:allb/fans

For categorisation of stops, you can have a look here :
http://reprap.org/wiki/Safety#Emergency_Shutdown_.28ESD.29_button

However, the demand was not for an emergency shutdown, but for something which may save some printer parts and notably the bed. Indeed, if you have a surface like a Buildtak, maintaining hot end pressed on it will make a nice conical hole.
To lift the head, a macro could be run from pin signal, but as underlined by David, we have to wait an issue which gives higher priority to user pin signal.

If there is nothing connected on the 5V input, the fact the the external jumper is set will not do anything. This may be bad if there is another supply.

About the USB connection, when starting the Fisher I asked confirmation (to RRP) that it was ok to have simultaneous 5V supply from USB and internal source, and they confirmed it (prior this answer, I was removing all 5V jumpers when using USB).
I also commented to T3P3 that this shall be in documentation when buying the Duet 0.8.5 (in dec 2015).

About the steppers, you can turn them on and off by G-code anytime without need to set them idling first.
You can create a cancel.g macro which will be executed when you cancel a print instead of resuming it, see useful macros: http://otocoup.com/aql/#hlp:allb/macros

You can find some info about fans (all Duets) here : http://otocoup.com/aql/#hlp:allb/fans

AussiePhil, that is a bit off-topic, but could you explain a bit (preferably in another thread) how you did cool your board, as the value of 1600 is indeed quite high for the Duet 0.8.5. Your machine is quite on the edge, it might be interesting to create a page with some specifications on the RepRap Wiki.

As for the value exceeding limits in the configuration tool, instead of enforcing values, may be a warning might be better or the possibility to overpass the told limit.

Out of curiosity, I tested R/C servo operation with a controlled fan output. Good news, it works !
That will allow allow any servo controlled sensor to work, including BL touch.
The command M106 exists and allow frequency control, so everything was available.
For operation from a pin on the expansion connector other than the fan control pin, we will need some G-code extension.
I tested on a naked Duet WiFi prototype, only supplied by USB, with 1.15 firmware and a very small servo (metal gears, preferable for this kind of experience).
The fan output drain current through a resistor (# to 470 ohm) supplied in 5V, so signal is 5V minus diode forward voltage.
Servo power is also 5V. I selected 5V for the fan voltage, but if you use Vin as fan voltage, you may add a diode to protect from voltage return through the fan LED (at least for Duet 0.8.5 and Duet 0.6, I did not checked the Duet WiFi schematic). Note my test was done without protecting diode (no need when only 5V on board).
You will find details, schematics and G-code command here : http://otocoup.com/aql/#hlp/connecting_servo

I made a small film of the test here : http://otocoup.com/aql/h/duet_servo_control.mp4 (size 11 Mb). The image is not stable, but this was more explicit than other films I done with a supporting tripod.

About the BL touch, Have a look at this thread : http://forums.reprap.org/read.php?416,669287,669287#msg-669287
Short answer : Not yet
Longer answer: It may be workable to output a servo compatible signal with a fan/signal output, using M106 code to adjust the PWM, but this is not yet tested.

The M206 can be used to do 'baby stepping', but as it don't return its status, it is difficult to handle it.

I described a simple switch based filament detection here :
http://reprap.org/wiki/RepRap_Firmware_macros#Set_of_macros_for_filament_detection

I published details and photos here: http://forums.reprap.org/read.php?416,668887,668887#msg-668887
However, I have yet to publish the stl and scad for the modified tensioner

I second this demand, as for miscellaneous reasons, I have done the calibration many, many times.
What I have done yet to help is making a macro for each sensing point. This macro go to the given point 0.5mm above 0, so P0 go to point 0 X, Y , Z=0.5, P1 go to P1 point X,Y, Z= 0.5, etc. This help a lot compared to the manual move to each point.