every week or so, i read a couple topics about people complaining that Blizzard doesn't have enough devs, or that it has bad devs. in both cases, those people suggest Blizzard should hire new devs. this is totally counterproductive.
a bit about myself in the hopes i give the impression what i'm talking about: i recently graduated from an IT college. i got education in at least 15 programming or software "languages", depending on what you call a different language. i have created 5 distinct projects for different classes, both alone and in group. i've had several 12-24 week courses on fringe matters related to software development and IT best practices, including moral issues, project management, IT service management and system development (deciding what to add to the system with the customer). i don't want to say that i'm the best man to explain this, but i think that compared to about half of the posters on this site (who have no IT background at all and don't even know Java from .net, and more importantly have not had to deal with a group IT project before) i at least have a reasonable understanding this.
now, there are a number of reasons why blizzard can't add extra devs that easily. i'll highlight the biggest here. the main background for this is a book titled "The mythical man-month", written by Fred Brooks. Brooks was a project manager at IBM and, through experience, laid the groundwork for software project management in his book. the book is known mainly for 3 main points, and a half dozen smaller ones. the biggest one of these main points, though, was what would be known as Brooks law: "adding manpower to a late software project makes it later".
this law might seem strange, but it basically is true because a project is often not easily dividable in separate parts. teams need to coordinate their efforts, and communicate with each other. this communication takes up time, and at a certain point, extra manpower means that the time saved by faster development is counteracted by the extra overhead from communication.
a quick napkin math example. suppose you got a game with 10 worlds and 5 levels per world, but a storyline like the witcher 2, where nearly every level effects each other. each level takes 5 days to design and code. if you got 10 designers on the game who each take 1 world (so 5 levels), each designer can work on the effect in their own world, and communicates with the other 9 designers for the effects. that's 45 channels in total. if every channel also takes half a day per level to discuss the effects of that level on the levels of the other person (so 5 levels per channel, since reverse effects would be nearly pointless), that's 25 days from developing the worlds (since all 10 developers can work alongside eachother), and 112 days from the communication overhead, or 137 days (4 months and 17 days).
now, if you get 50 developers who each take 1 level, you only need 5 days to code it. however, 50 people means 1225 couples of people who need to communicate. again, if each level takes half a day to discuss (1 level per person), that's 612 days of overhead, for a total of 617 days of work (20 months and 17 days).
now, if you look at it from a blizzard point of view, it holds very much true. consider a fight like the Lich king in ICC:
1) art designers for the mobs (lich king, val'kyr, ghouls, horrors, spirits, terenas, spirit warden. that's 7 mobs to design).
2) animation designers for each of these mobs. again, 7 mobs to design animations for. you also notice that each of these mobs needs communication between the art and animation designers, because things like clipping are ugly.
3) spell effect designers for each mob. each spell needs discussion with the art designers and the animation designers. also, some spells need discussions between multiple designers (like the ghoul and horror summoning). i've counted 24 spells in the fight.
4) environment design. environment needs to be to scale with the mobs and the spells, and things like remorseless winter effect the environment (exploding pillars).
5) doodad design. stuff like the disappearing platforms.
6) encounter design. calculating numbers and such.
7) scripting of encounter, mobs and spells.
so you got 7 aspects, taken by at least 7 developers. that's 21 channels. if you got separate designers for each mob or spell for the first 3 categories, and separate designers for each type of scripting for the last category, that's 44 people, with worst case scenario of 946 channels. i'm not going to perform math on this, but you see that's it's pretty similar to the 10 worlds example of earlier.
now, i simplified this a lot. also, there's the conception that it's not just the quantity but also the quality of content that is under dispute. that's not something that will be solved by a simple changing of the guards. the people at the top have an idea of how this game should progress, and they won't easily be swayed to change that. new developers won't magically change that.