The problem with a 'one book to rule them all' approach for 40K is almost exactly the same as the problems faced by White Wolf in writing scion.
The difference in power level and type of story told between an inquisitorial henchman and a space wolf champion are so markedly different that creating a system that functions for both is nearly impossible, much as the difference between a hero and a god in scion.
The task of making a system that functions for the lowest henchmen and the mightiest of the imperiums heros is made harder still by the use of a percentile system.
Add to this that a 'One book' would need to include, comprehensive vehicle rules, space flight , zero and vaccum operations rules, warp navigation, capital ship combat, aerial and space dog fight rules, psychic and sorcerery rules for all levels of power and an armory containing everything from shivs to titan weaponry. In short, you would end up with a book the thickness of dark heresy, without a setting.
If someone went for that approach, we almost certainly would not have it in our hands now, and it would likely be a considerably worse product. It makes considerably more sense to focus on specific areas of interest, and produce a system, which reflects the spirit of that style of setting.
Not necessarily. I don't need space flight and capital ship combat in a Dark Heresy campaign. That would fall under specific rules for Rogue Trader. What I need is a list of skills that all settings have in common - character generation; individual combat; common skills; common talents; common weapons; common body replacements; Insanity; Corruption; how to and when to use skill checks; basic background on the 40K universe; etc. It could actually be a smaller book. Then you have settings book focus on Dark Heresy, Rogue Trader, Death Watch focus - give the background of that specific setting; professions, skills, talents, traits common to the setting; specific character generation differences; powers like psionics and sorcery; etc.
In my opinion, Scion screwed things up the way they did it. The jump in power levels between the books was just odd. It caused problems in our group because some people wanted to move up the next power level as soon as they could so saved up all their points to do that while others were buying useful skills, etc. before moving up. They should have made it a smooth scale instead of one that jumps in power level so drastically. It was rough for good characters who didn't save up their points who were still at the hero level to be effective when others were already at the demi-god level and so much more effective because they got that huge bump up in characteristics and skills. It made for an unbalanced game at the meet points between the three books. We eventually abandoned the game because of these flaws. That and having to roll unreasonable numbers of d10 as you go up in power.
Actually, I think in some ways percentile systems do pretty well with henchmen because you have a much wider range of skill levels to use. You can give the common person low skills, henchmen higher skills, competitent people higher skills, and scale all the way up to a mighty hero of the Imperium having skill levls in the 80+ range in things they are extremely competitent in. Granted, that is where I see that DH falls down though because Acolytes that should be at least somewhat compitent even at introductory levels, aren't. Once they get a few thousand experience under their belts, they start to be at the level I think they should be at in the beginning.
Oh well. No game system is perfect. But as long as I enjoy the setting and the rules don't get into they way too much, then I'll play it. And while I see problems with the DH rules, they usually don't get in the way except at the early levels and so as long as the GM handles the new acolytes with kit gloves, they can survive long enough to become competitent enough to make the game enjoyable and get emmersed in the system.