Negative things I've learned from Hedgerow Hall:

* Removing death from your game doesn't remove player killing; it simply changes the nature. Players will always find a way to "kill" other characters. In Hedgerow Hall's case, player killers would continue to attack characters they wanted "dead", thereby keeping them unconscious until the player logged out from frustration.

* Removing player killing opens up many other aspects for abuse, including the ability to steal people's in-game items with no real consequences. Any attempt to deter them from this through in-character means will be met with derision and a generally out-of-character attitude. ("Ha ha ha! You can't get it back, ever! Now I'm going to go around and impersonate your character, just to tick you off!")

* A complicated philosophy is often lost on an open community. Players do not appreciate strange or groundbreaking things -- they prefer environments which are more like real society and therefore more comfortable for them to understand.

* No matter what precautions administrators take, there will always be griefers, and there will always be players who believe that the administrators are out to get them. This is an unhealthy side effect of enforcing rules strictly; rather than learning not to break the rules, they simply believe that the rules are being unfairly applied. The anonymous nature of the internet doesn't help any -- the general public has no way to corroborate text to actual expressions, so the mood of the administrator is not obvious to them (and is thus probably perceived as hostile given the nature of the actions).

* If players can abuse a bug or a missing feature, they will do so, regardless of any statements against it, and even in defiance of any administrator action promised or performed. There are only three solutions: 1) shut down the game until the bug is fixed; 2) permanently ban all such offenders as they become known; and/or 3) provide rigorous testing and sanity checks within your code so exploits are visible right away.

* Players will very rarely report bugs unless they personally stand to lose as a result of the bug remaining unseen.

* Opening a game to public consumption is both rewarding and stressful. You must weigh your personal taste for the former and distaste for the latter to determine whether you will be enriched or depleted from the experience.

* Very few people know how to roleplay properly. That is, they know how to "play a role", but the lack any foresight or knowledge in regards to how to "play a role believably". This doubtlessly will result in players claiming absurdities or impossibilities as backgrounds and abilities, and believing they are perfectly acceptable.

* If you are making a roleplay-enforced game, you will inevitably attain people who neither know how nor want to know how to roleplay, yet who want to continue to play your game as "just a game". Attempting to teach these people how to roleplay will generally result in frustration on their part as well as yours, and will often breed a dual nature -- where they will roleplay only if they perceive they are being watched by an administrator, and cease to roleplay at all other times.

* Any editing of old features -- generally for game balance -- will usually be met with skepticism. Any new features often will be met with much appreciation, even if they upset the game balance.

* It is extremely easy to ignore a bug in favour of adding a new feature.