Awesome movie, loved it.  Definitely a my generation thing.

In a word, no.  You can try quitting before the server crashes which will save your expereince (afaik), but thats a pain.  You'll just have to bear it for now until the server is a bit more stable- there are a few bugs with the new release and all.

The simple fact is developers are always in short supply.  With such an ambitious project as an mmorpg (undoubtly one of the hardest things to develop well) the planeshift project needs very high quality members.  You're essentially asking people to do stuff which they could be (or often are as part of their day jobs) paid good money to do, in their spare time in between everything else that they have going on in their life.
In addition to this, there is already a base of very good developers at the top of the chain in Planeshift who have much of it planned out, and who make alot of desicions about it- these means new developers who come in don't really have much control of anything they're doing for quite a while (in a design sense).  Thats just part and parcel of such a large and complex project, but it means developers have to have a real passion for simply coding and letting others make desicions about where things are heading.  Alot of coding that developers do in their free time is stuff they do because they want to experiment and have control over the design process, but its simply not feasible to let people coming into the project have that sort of control in Planeshift.  So its one thing to have a developer who likes messing about and coding stuff up in their free time, and having a developer who is willing to simply debug a complex system for hours on end and fix problems that come up.
The attitude of "We have this game, why aren't lots of people volunteering to work on" shows a lack of understanding of software development.  Its such an arduous, complex, time consuming and (at times) frustrating task that I'm honestly suprised at how far it's got so far and the passion of the people who work on it.
The only thing I can think of that would encourage more developers to work on the project would simply be better documentation.  To be brutally honest, the current code documentation of planeshift lies between non existant and shockingly bad.  But you know, its a case of being one thing to pinpoint something lacking, another thing to fix it- every developer could stop work for a month and work on documentation in that time, but then the players don't get any new developments for an entire month. 
Simply put, the inherent problems with attracting developers to work for free to a project are far bigger than any problems caused by <insert conspiracy here>.  Its just a fact of life.  If everyone that got annoyed by the pace of development of Planeshift went away and learnt c++, and if only one tenth of those people had the intelligence and skill to contribute in some way to the project, It would probably be finished by now.  Until then, when the small group of people who are developers and spend time working on the project see posts like that, do you really think much good is going to come of it?

There are already measures in place to prevent 'speed cheating' such as running faster than is allowed, etc.  I'm not sure whether these options are currently turned on in the server and being monitored by somebody, but its definately in the code.  If you have the name or a screenshot of who was doing this you could contact a high level GM about it.

What kind of loser do you have to be to cheat in a game that has no winner?

Granted, none of us would think to do this, but that doesn't solve the problem.  Its a fact of online games that people will try to cheat.

In addition, I'd prefer to base the chance to hit on the character's actual stats and levels.  But it appears most people disagree with me.  The few I have participated in turned into - who ever rolled higher gets the hit (better than nothing).  I guess most people don't want to give out their actual stats and levels, and then agreeing on what each would need to hit/block the other is a whole other set of problems (and probably alot of OOC discussion that would kill the mood).

The entire point of having PS on a computer is that the system simulates all this for you, so you don't have to worry about it.  Granted, there are some problems with duels at the moment due to very powerful weapons and what not, but if people are still resorting to rolling dice in PS for duels in a few years time, then something has really gone horribly wrong.

How do you propose you would get them back?  By clicking on the appropriate icon (in the case of stats window, inventory, etc)?   The alternative being they fade out so you can see through them, but remain there so you can get them back by simply mousing over the area where they were... but I suppose that could get annoying.
So basically the windows would toggle off after a set time of inactivity, and then toggle back on when activated.
I think its a good idea, because the interface does really get very cluttered at the moment.

This thread is a month old, so perhaps the person who asked the question isn't still around, but I'm going to respond anyway for reference if there are other people wondering.

The npcclient Controls all of the actions of the NPCs in game.  It is a program separate to the server that receives information about each NPC in game, and then makes desicons about what they should do, and proceeds to send the appropriate network messages back to the server to control the actions of the NPCs.

Each NPC contains a list of behaviors, which are actions it can 'do'.  Each behavior will do some processing and then send a network message to the server to execute that behavior- things like move, start attacking, turn around, pick up and item, etc.  This is the same way that the psclient controls it's own character, by sending these network messages to the server.  Each behavior has it's own class (MoveOperation, MeleeOperation, etc) which inherits from the base class Behavior.  Each behavior has an overloaded method Run() which will execute the code necessary to 'run' this behavior, as I mentioned above.  There are also some other methods like interrupt, etc.  Each behavior also has a 'delta', which is how much the NPC wants to do that action.  This is how the NPC decides what to do.  It iterates through all it's assigned behaviors each timestep, and looks for the one with the highest delta, and then runs it.

The two other bits of the AI are perceptions and reactions.  A perception is something that happens to an NPC; it 'perceives' something about its environment.  For this to happen the server must send messages to the npcclient whenever something happens that an NPC would be interested in, and then npcclient will trigger the perception on the appropriate NPC.
Lastly, reactions 'react' to perceptions.  A reaction specifies what perception it is to react to, and what behavior's delta it should modify as a result.  Thus, whenever a perception is triggered on an NPC, it will loop through each of its reactions, and see whether that reaction is for that perception.  If it is, it will modify the delta of the associated behavior accordingly.

To run through an example:
A player attacks an NPC. 
On the server:
The server generates an 'NPC attacked' message with appropriate information, and dispatches it to NPC client.
On NPC client:
npcclient receives NPC attacked message, and matches up which NPC it is for from the NPC id.
npcclient then triggers a perception of type "attacked" on the correct NPC.
The npc iterates through it's list of reactions, and finds which one(s) respond to being attacked.
It finds a reaction to being attacked, which is to add 400 delta to the "fight" behavior, which is does.  Fight behavior now takes over the "do nothing" behavior as the highest delta.
On the next step through all the NPCs, "fight" behavior (which is a MeleeOperation) will be the highest delta, and its run method will be executed.
MeleeOperation->Run() does some processing and if the conditions are correct dispatches a StartAttack message to the Server.
On the server:
Server receives startattack message for that NPC, and does the server side processing of attacking (calculating damage, when it can attack, etc).

Thus, to write an AI using this framework, you need to specify Behaviors and Reactions.  These are specified in the npcbehave.xml file, which is then loaded in by npcclient.  The methods to load these in and convert them to valid Behavior and Reaction objects are handled by /src/npcclient/npcbehave.cpp and /src/npcclient/perceptions.cpp.

Hope this gives you a general overview.  I'd strongly encourage you to download the npcclient client and have a look around if you're interested, as its quite an interesting part of planeshift, and a part that will definitely need committed contributers for in the future.  There are lots of enchantments that you could add if you know how to code c++.
Some of the specifics I mentioned above might not be 100% correct (like the name of some methods, and the names of some network messages as I didnt have the code open at the time) but it should give you a general idea of how it all works, and you can look it up yourself if you'd like to.

Alright, but can we be sure they really realise it? Everyone knows there is a problem, but where is it? Which part? I believe these discussions help to understand nature of the problem. Personally I'm suprised, that for example the idea of NPC asking you what you mean, basing on keywords is over 6 months old and we can't see it in game. What we see is bunch of new quests, NPC answers. Only what for all this if you can't dig that from the NPC?

As I said before, if you think that constructive points are be raised and it will help the developers, then make a wishlist thread detailing the outcome.  That will be helpful.

I don't think the idea of basing keywords if merely 6 months old- its a simple concept that has surely been around since planeshift started.  The point which you don't seem to get by expecting it to be in game is that obviously all the developers are busy with other things at the moment.  I'm almost positive that new quests and NPC answers are not done by someone on the engine team who would be otherwise coding something.  That is why we have new content but not new developments in that area.  It's better than nothing.  Until then, construct away, but don't be surprised when an idea is only 6 months old and yet "we still haven't seen it in game".

Who says this idea is so great anyway?? In theory it works well, but in theory the current system works good anyway.  I am in no way surprised that this isn't the biggest priority at the moment because it would have to be prototyped and extensively tested first.  If you could come up with an algorithm for describing in detail how the system will go from the player's input to the NPC reaction, that would be helpful.  Describing how the system gets it done is alot better than providing examples of what it does.

Lastly you have to accept the priorities of the developers- this may seem the most important thing to you, but other people obviously have other priorities- there are often more critical bugs to be fixed.  Even if you could describe the algorithm complete with system flowcharts and input-output diagrams it could well be longer than 6 months before it's in game.