MikiSJ: I and many others here have found Dr. Glover's lecture really informative. Others have pointed out that it doesn't necessarily provide the "optimal" subexposure length, as it does the shortest, much below which you are beginning to add unacceptable levels of noise. Once you have chosen the total integration time that you want for a given object, you definitely can break that up into subexposures set by his formula. But there is no "harm" in going longer, it's just that in doing so you will need more precise tracking, perhaps a better mount, guiding where you might not have otherwise had to, etc. Longer exposures also increase the risk of imaging time lost due to a cloud entering the field. That is a major bugbear at my VI observing site, where small, rapidly moving clouds are the norm, and cloudless nights rare. Satellites or planes can be handled with the right algorithm, but their odds also go up with exposure time. So there are distinct trade-offs in going beyond his recommendation for subexposure lengths. It certainly is valuable to know how long is long enough, and in many cases, that may well be shorter than tradition or habit might suggest.
His approach boils down to choosing subexposure lengths equal to (or longer than) ten times the square of the read noise divided by the sky electron rate. Your camera specs will include the latter; while some of Dr. Glover's tables provide values for the latter, it's far better to use his very helpful web calculator:
Note that you can directly enter your specific values in each field, to a level of precision higher than that allowed by his increment and decrement buttons.
The above approach is aimed at faint, extended targets such as nebula. Star clusters are another thing altogether, and there you may need to use shorter subs, to avoid saturating stars and losing their color information. Subexposure lengths for these targets are best determined empirically.
All the best,