I thought I understood the principle of the K metering system (I use the more conventional Peak/RMS type metering), but recently I've been reading some stuff which makes no sense to me whatsoever.

The stuff in question talks about calibrating your monitors such that a specific SPL level is achieved when running specific test tones through your system.

Eh?  I cannot understand how the output level of your monitors has any bearing on the level of signal being monitored by a master bus K-system meter in a DAW.  So let's say I've configured my system like this, but then:

a) Move slightly further away from the monitors
b) Put in some earplugs
c) Turn the monitors down (or up)
d) Use some headphones instead

...etc.  How does any of these actions invalidate (or affect in any way) the metering which is occurring inside the DAW?

Quite frankly, I am left completely baffled by this requirement.  Perhaps it isn't really one?

The volume knob alone is huge... What "ruler" is supposed to be used to calibrate with?

On the cell here, but the OP sounds like the article is amiss on what/how calibrations are done. Is there a link to what you were reading?

I can't remember the link to the original article I read a few days ago, but the internet's awash with them.  Here are the first few links from a gooflz "k system monitor calibration" search:
 
https://www.gearslutz.com...fication-k-system.html
 
https://www.gearslutz.com...ibration-bob-katz.html

http://forum.cockos.com/showthread.php?t=76587

http://www.meterplugs.com/kmeter

K System is about both.  It is about the levels you are working with in terms of the rms component and the digital ref level that represents your 0 dB VU level and all inside your DAW.  Many digital systems use -18 dB FS as a level that can represent 0 dB VU.  K System has three ref levels to work with and in my opinion they are well chosen. K-12, K-14 and K-20.
 
It is good to work at a common reference SPL level in your room as well.  At certain SPL levels our ears offer a flatter more even frequency response.  (Not perfectly flat but close to in a weird sort of way)  There can be discussion as to what a good SPL level might be.  I agree with Bob Katz in that 83 to 85 dB SPL  (C Weighted) is a great level for me to work at here in my studio.  (It satisfies OH&S requirements eg for 8 hour period)  It is good to monitor at a range of SPL levels in your space.
 
Your monitoring SPL level is tied to the ref level inside your system.  If you calibrate your system to K-20 and make that produce 85 db SPL level in your room then you are set.  But if you decide to work at K-14 instead then the SPL levels in your room will suddenly get louder and now be at 91 dB SPL which is loud.  So you need to knock back your monitoring level by 6 dB back to 85 dB SPL again.
 
There is more too.  About using VU meters and keeping all rms components consistent right throughout your DAW.  It greatly helps tracking because all your tracks will end up at perfect and constant rms levels.  People sweat the peaks way too much. They take care of themselves and won’t clip anywhere usually when done right.  It allows for correct ref rms levels on all your busses and finally your stereo mix too.  It makes mastering much easier and it all works perfectly with Loudness metering.
 

 
This may help clear up all your confusion:
 
http://www.soundonsound.com/sos/may14/articles/reference-monitoring.htm

ston
a) Move slightly further away from the monitors
b) Put in some earplugs
c) Turn the monitors down (or up)
d) Use some headphones instead

These are indeed 4 quick ways to completely defeat the K-system. That's because the system is about listening levels first, and absolute levels (as shown in your meters) second. The idea is to establish a consistent correlation between output level on the meters and the volume you hear through your speakers. That's why you calibrate the speakers.
 
The levels suggested for calibration are actually a little too loud for most people, so even those who follow the K-system guidelines will normally monitor at a lower volume. That's actually OK, as long as you're consistent. But as soon as you put on the headphones (or earplugs) you're no longer using the K-system as intended. At that point you're just using one of the k-system recommended RMS values as a visual target.
 

I might not be doing this correctly...but when I mix I usually use K-14 and I listen at a minimal level. When I perform a mixdown to distribute I usually export at a higher level but still using K-14...my exported mix usually stands up well on most systems with half volume being at comfortable listening level and headroom for a lot more if desired. I just had the pleasure of talking with the guy who invented the K system   but I didn't think to ask him about that :)

Thanks for the replies folks!  I understand what you're saying, but fundamentally disagree for a number of reasons.
 
I shall continue to use my secret weapon when monitoring (this is also the same weapon which was used to end the loudness wars!):
 

i think you've missed an important element of this discussion.
 
even tho the 83 db (85) setting for calibrating your monitoring station may seem a bit loud...
 
it IS the level you need to meet the human condition- meets - munson-fletcher curve -  meets translation across multiple platforms.

^^^^
 
After seeing this in more detail, the monitor portion is really more for consistency (as bitflipper mentioned). Even something as simply as the same volume knob setting (including position, etc.) for a given digital out will yield more uniform results. There are free apps for cells that will sound meter sufficiently to choose an adequate volume placement, etc.
 
Another factor which doesn't get mentioned as much as I feel it should is fatigue. Coming from a production environment, there is only so much you can push the human body before fatigue (errors) begin to appear. Not taking breaks, not enough adequate rest, distracted (or even overly focused), throw a whole gambit of "human error" into equations even with the most precise tools and working conditions available (i.e. identical environment).