And I think alot of people, including Jinx, objected to the way in which this is implemented in that thread - with DG-specific items. The intention remains the same from both sides, though, which should be encouraging.
So what are some other ways we can create deep gameplay using DoTA's lessons, in Demigod?
There's a lot of things in my above post that can be translated to demigod if they really tweak their numbers. Creep last hits and creep denies are easy to implement and are highly rewarding when you get good at it.
They also need to make sure that you never, ever fight with the UI. The UI should honestly feel like your brain is directly commanding it. There shouldn't be a "translation layer" where you have to think what you want to do then think about how to do it. It's really hard to put that part into perspective right now because the current Beta's UI is absolutely a train wreck. Beta2 promises to fix some of the shortcomings so we'll see.
They also need to implement things that can help you have better map awareness. Heroes should have an outline or something to help you distinguish them from the creeps in the battlefield. On the minimap (which is coming in beta 2), they need to have large, clear icons telling a)what team they're on and b ) which player they are and c ) what hero they are (this one is optional). I'm 99% sure this will at least be implemented, since it's such a basic concept.
They can also implement features that benefit from twitch skills. Even Dota benefits from good apm and demigod should be no different - especially if they want to attract competitive gamers.
Another critical thing they need is to deal with momentum. If a team has a strong momentum and is doing well, they shouldn't automatically win because of it. The losing team needs a way to fight back otherwise there will never be a point in playing a game to completion.
Hero-wise, I honestly can't really see it get to the level of dota with the current information they've given us. It may be a preference thing, but I really think that 8 heroes, each with ~8 skills, will get stale pretty quickly. If they can somehow keep it always feeling fresh and really ensure that no game plays the same, then i'll be happily wrong about this.
Now, going back to the items. Item depth can also really help the game's level of depth. However, if the items degenerate into more-or-less of a case of rock-paper-scissors, then you're actually taking several steps backwards, imo. Personally, I liked the recipes from DoTA but i can see how they'd pose a barrier to new players. That said, GPG/stardock needs to come up with something that makes its item system a system that warrants some degree of studying and experimentation.
I'd also like to make a point that gpg/stardock needs to beware adding things for the sake of adding it. When you shovel features into a game that aren't really related, or don't compliment the central gameplay, the game will end up feeling like a smorgasboard of fluff. The most critical thing is to ensure that the foudation of the game is rock solid before adding stuff. They need to make sure that even at "level zero", without the distracting additions, the game still has many degrees of complexity.
Lastly, they also need to REALLY beware adding "cop-out" features. Let's suppose there's a feature that's suppose to either solve a design problem or introduce a layer of strategy. However, when the designer can't really think of a good way to intrinsicly make the feature a part of the game and instead unelegantly slaps it on, that's a cop-out feature. A good example of this is when there's something intrinsicly broken with the game and instead of correcting the problem at its roots (that might mean tweaking the engine itself or its battle system), they throw on a band-aid to 'patch it up'. The problem hasn't gone away, it's just not as obvious. When people get to the "pro level" the flaw will become significantly more visible.