In my recent years playing MMOs, from Halo to the Madden series, I have came to realize that after looking at all the great things they have available, from many levels and and great graphics, I have come to believe that maybe we aren't ready to provide such detailed graphics and loads of options to help further gameplay without taking a huge load off of speed and reaction time between the serve and the player themselves. Technology has grown to where we can easily take use of these more advanced methods of making games look better and to keep good gameplay without slow downs, but with these games having to deploy many servers to compensate for handling processing millions of pixels and commands sent by each player, then you come to wonder if they are really neccessary at all. MMORPGs have to place more restrictions due to many having to cut dramatically down on graphics in order to keep players happy gameplay wise, due to expecting a large number of players. Unlike the normal sport or first person shooter games that demand lots of graphics and gameplay options, MMORPGs are simply a world that changes state and require the player to simply progress a pixel representaion of a character. And from recent studies, people aren't so much concerned about it. As long as it is fair and decent. Load times are much smaller on MMORPGs than on the other types of MMOs after looking at this, and it is being concluded that maybe every MMO developer needs to take into account that despite our advancement, we have yet to come up with a diehard solution for allowing both high level graphics, high level gameplay, and high level speed all at once, and to make us all happy, a simple standard should be placed to the limits of the game's specifications in order to provide the best speed and least loadtimes possible. A balance can be set that will assure all of what a designer wishes, but just doing this will require sacrifacing of extreme costly game functions. I will like your opinions on this and maybe some controversaries. Should speed be taken into recognition with reducing gamer's expectations in terms of other design concepts aside from high speed, for we know even a single lag can be the difference between winning or losing, and in term, if you're only taking one lag or an entire internet log off due to excessive information being passed all at once?

Quote:Original post by dmreichard
It sounds to me as if you believe every single pixel is sent across the internet, which is false. Any lag has to do with the player's internet connection as well, not with their system power in most cases. There is a fine line between net lag and low FPS.

Ah yes, I would also like to add that there is never any graphics processing being done server side, and successful MMORPGs are hosted on extremely powerful machines with insane amounts of bandwidth. I believe you are mistaken to presume that servers can not keep up.

I think you're missing the point. There are inconsistencies in his post, but what he's saying is true. MMO's sacrifice lots of things to be MMO. I think he's suggesting that we should perhaps stray away from twitch gaming in MMO's because of the technological limitations.

Twitch gaming in MMO's has two major problems.

The first is net lag. This won't go away. There will always be latency / reliability issues with internet traffic to remote servers. This is usally remedied by using coarser timesteps in simulation, a very tight datagram design and a variety of extrapolation, interpolation and predictive methods. It's not really possible to get around 'linkdeath / lag kills' because there's no difference (as far as the server can tell) between a genuine linkdeath (failure somewhere between server and client) and a client pulling their modem lead 'cos they're losing.

The second problem is one of simulation granularity - complex physical calculations with many moving objects or state-dependent objects (like players) require a fair whack of processing to resolve - too complex a simulation occupies more cpu-per-player than is worthwhile, and can make the overall architecture unwieldy if results have to be distributed across a cluster. As processing power increases along with (internal Lan) bandwidth, coupled with the availability of discrete physics accelerators, we'll start to see more complex simulations being viable for MMOs. Although I've concentrated on physics (more appropriate to the general school of twitch games), this applies equally to complex state machine based games (very complex RPGs), only without the extra help from physics coprocessors.

Load time is purely a factor of the hardware involved (RAM drives vs. traditional HDD), number and complexity of assets, and frequently incorporates a decompression / setup stage, where the locally cached asset is unpacked / generated from appropriate source data (eg terrain from heightmap). Good designs for MMOs tend to ensure that any distributable asset is distributed before the system goes live (servers come up, client connects).