I have perennially tried to take the position that rendering is, or at least should be, a client-side problem.

The most powerful graphics processor in the known universe is your own imagination.

Your favourite "artist's impression" of this that or the other thing may well not be my favourite "artist's impression" of that very same thing.

You might be colourblind to the very colours I try to force your visual display device to display. In fact you might even be completely blind. So there is little point in my resorting to pictures at all unless there is no other way to convey to you what exists what is available what can be done what you have what you can do and so on unless I am confident that you can actually see pictures and that the pictures do in fact convey the pertinent facts without distracting from the facts with a bunch of irrelevant visual clutter.

One of my ideal or dream game-clients for people who like eye-candy would be a text to eye-candy translator that can render pretty much any ordinary item of literature such as a Dickens novel or a Shakespeare play directly from the text.

Such a client would allow game designers to focus on the game instead of on how various species using various sense-organs might like to represent the game to their particular combination of senses or might find it convenient to represent particular aspects of the game upon which they might choose to focus at any particular moment.

I have not spent much time in "Second Life", largely because until recently my usually years out of date personal computers were unable to render it. Now I have one that can render it so I am able to find that I am not at all impressed with what it is rendering.

(Actually, technically I still cannot be sure whether maybe it is merely the rendering that is unimpressive and by leading me to think what I am seeing is what the underlying models on the server intended the rendering to look like it is leading me to think the world being rendered is unimpressive.)

What is unimpressive about it is all the sheer hassle pain expense etc etc etc going on and on and on about the importance of "realistic rendering" and "eye candy" and so on, all the expense of upgrading year after year after year to try to stay as few years behind the high end state of the art graphics cards and the net result is this?!?!?! Sheesh! It is not eye candy at all! It is cartoony! Why even bother? What is the big deal? It is a gross waste of my time to watch let alone a gross waste of resources to render.

But one important thing I do notice is players can provide their own appearance if they choose. They can pay artists to provide them with different appearances. Aha! Could it be, then, that I could fob off the rendering problem onto player-directed / player-employed third parties? Actually achieve the "ideal" of the players choosing for themselves what they want the game to look like and sound like, freeing me to design the game itself instead of having to worry about whether any given player's favourite artist's impression of it will agree with my favourite artist's artist's-impression of it?

Are you looking for some type of feedback about your post or are you making a statement something you would like to talk about of your post? Not sure what you are getting at.

Good point! Thank you.

I suspect the underlying question is probably whether it is in fact necessary to drown a game in eye candy and ear candy in order for it to be viable.

Getting specific, I have so far been using three existing engines to attempt to render various aspects of my game, and historically, even dating back to the use of off the shelf miniatures (albeit sometimes at least some of them hand-custom-painted) for tabletop play, it was always played using "found art" / "found objects" / "off the shelf props". Basically we were never producing a movie a novel could be written of, we were, rather, producing novels or plays that some day some movie producer could in principle make movies of.

So maybe another likely underlying question is whether literacy is headed toward becoming as outre as being able to hear a symphony in your head by eyeballing the grooves of an "LP"?

(LP: an ancient audio-recording media, somewhat akin to stone tablets...)

Good insight! Thanks!

In fact some of our designs had in mind simply having indexed fields and items and so on with the ideal intended being that each player could label those fields or items however they chose, so one player might think he was a cowboy wielding a sixgun while another might think he was a spaceship equipped with a cannon and six cannonballs...

To make that work though we would have had to eliminate player-provided chat text, so that each utterance that could be made could also be similarly translated between "skins".

(Otherwise trying to chat to each other could make it very swiftly evident you were each in a totally different milieu.)

Part of the motivation back then was to maximise the use of a limited pool of players by allowing all available players to all be part of the player pool of all available milieus without any of them having to play any milieu other than their favourite milieu.

I am currently basically working with what is freely available, so for example I have not included nations into my Galactic Ruleset for Freeciv, instead I assume either people are likening any civilisation they encounter to the most similar nation they know of from the widespread legends / myths of a planet where all nations of humans all existed ("Earth", an obviously fictional planet since we can plainly see in the "real galaxy" that even just a few nations on one planet always seems to lead to war until one nation has the planet to itself, the others either wiped out or assimilated or flown off to another starsystem in {|a} spaceship{|s}) or the civilisation is actually inspired by those legends/myths and maybe the promulgator of them with respect to their specific nation's mythology / creation myth or whatever. (Once upon a time our ancestors came from the mythical planet named Earth, and here are tales of what our still existing nation on that pie in the sky planet have been up to lately even maybe...)

The fact that players of Second Life actually themselves pretty much directly finance the development of graphics though leads me to hope that maybe some players might like to play a nation of their own design instead of being limited to only those nations that happen to have been included in the latest mainstream release of Freeciv. If so then maybe if they come up with nice ones with good shield and flag and description it might start being worthwhile to turn the Galactic Ruleset into a Galactic Modpack, that is, include graphics such as new nations as well as mere ruleset files. At that point, having broken the distinction between a mere ruleset and a modpack it might start seeming worthwhile to throw in more graphics too, such as new images for units and technologies that don't exist in the mainstream plain old Freeciv distribution instead of falling back to images from the default for all the new stuff.

(All the paratrooper type things, such as transporter-delivered redshirts etc, currently all just look like paratroopers, all the marine-derived things, like Mobile Infantry and Shielded Troops and Galactic Marines and so on, look like marines, gravtanks look like tanks, gravbarges look like transport ships etc etc currently. Partly this is almost necessitated by the fact that Freeciv has an ever growing collection of "tilesets" in various styles and for various topologies and sizes, so there would always probably be various "tilesets" out there that lack specific images for some units featured in some rulesets/modpacks.)

Quote:Original post by markm
I have not spent much time in "Second Life", largely because until recently my usually years out of date personal computers were unable to render it. Now I have one that can render it so I am able to find that I am not at all impressed with what it is rendering.

(Actually, technically I still cannot be sure whether maybe it is merely the rendering that is unimpressive and by leading me to think what I am seeing is what the underlying models on the server intended the rendering to look like it is leading me to think the world being rendered is unimpressive.)

Have a look at OpenSim, it is based on Second life, but you can run your own server that is indipendent of Second Life. Also, Look at Thrid Party Viewers for OpenSim and Second life.

As you indicate in your posts, you want to have the viewer indipendent from the source. This is exactly what OpenSim and Second Life have donw. The viewer is indipendent of the server and you can view the graphics with any compatable viewer.

As OpenSim and its viewers are opensource, anyone (with the right skills) can make a viewer and make it display anyway they want.

However, the problem with second life and opensim is that because the world is dynamic and user created, the tricks used by game developers to get the high quality graphics can be used easily in them. Often in games, to create good lighting effects and texturs, the texture for a specific location will have "baked" into it the effects of lighting in the area. That way the game's graphics engine does not have to render the light and calculate its effect on the texture, it has been done before the game was packaged and shipped to the stores.

This is why, in programs like second life and opensim, the graphics quality is no where near as good as in the latest games.

Quote:Original post by markm
What is unimpressive about it is all the sheer hassle pain expense etc etc etc going on and on and on about the importance of "realistic rendering" and "eye candy" and so on, all the expense of upgrading year after year after year to try to stay as few years behind the high end state of the art graphics cards and the net result is this?!?!?! Sheesh! It is not eye candy at all! It is cartoony! Why even bother? What is the big deal? It is a gross waste of my time to watch let alone a gross waste of resources to render.

The costs of rendering, in real time, all the lighting, bumpmapping, geometry and other layers of objects in a game is enourmous. Now even the latest graphics cards could do what we see in games in real time. Most of the rendering is done during the game development and it is an image of this render that ends up as the texture on objects in a game.

It is a cheat, and it is a good cheat because it allows games to give us far higher quality graphics than we could have had otherwise. It is only because in programs like Second Life that can't pre-render all this extra stuff that we end up with "cartoony" graphics.

If you remember the graphics quality a few years ago, this is about the quality that second life has (and at that time the graphics in those games were partially pre-rendered to get that quality back then). The reaosn that programs like second life have graphics qualities that lag behind current games is because programs like second life can use the cheats of partiall pre-rendered graphics (although with work you can do this a bit in second life).

Thanks! Good exposition.

Maybe that now brings me to a thought that came up while delving into the networking area of the forum and its FAQs and links therefrom and so on...

I would be just as happy to watch the movie later, after having already read the book or maybe it is more like having already written the book... in other words, caused the history, by means of playing the game, that the movie will be trying to depict when it does eventually come out.

Reading about the problems of synchronisation for twitch based MMO systems I found myself wondering why the heck tie the twitch to the movie instead of tying the twitch to a twitch-pad or something like that.

Certainly it does not offhand seem unreasnable to me that if I were a cowboy using a pair of sixshooters in a fastdraw type shootout I very well might not bring the sights of both guns up to my eyes to obtain a line of sight presumably from each eye to each target? My guns would be down nearer my waist maybe, likely not impinging in my field of view whatsoever.

So if it is desired to measure the human operator's reaction-time and pointer-positioning accuracy why do that in the middle of the movie or telly screen, getting in the way of watching the movie, instead of doing it off in a corner or off on a tabletop touchscreen separate from the movie screen where we can measure the player's speed and accuracy without worrying about whether we measured it before or after the movie about it gets shown?

Is it really that important to stick the twitching hand(s) up in the line of sight? Don't we really just need to know how many milliseconds it takes them to react to how many pixels that change how much in lumens hue intensity or whatnot and how long their hand-head seek takes to seek through x degrees? Then take that into account when making a movie about which character gets off more shots more accurately at targets consisting of x number of seconds of arc of their visual field?

Heck we could even have both players play reaction-time, observation and pointer-pointing tests in their offline time then use the data from those tests to create movies about "what would happen if" this one had a shootout with that one.

Then too, must the character always sleep when the player sleeps, be off doing something other than being shot at when the player is doing something other than being shot at and so on? We need the player's twitch-statistics anyway presumably in order to resolve things that happen while the player is away from the console!

Most of the time really I would probably prefer not to watch the movie until long after the story has been worked out, hopefully including whether what happened was in fact spectacular enough to be worth making a movie about, let alone watching a movie about...

Meanwhile though, my buddy's laptop makes the crystal space based virtual world look like pretty good eyecandy, much better than this deskptop I am using makes Second Life look...

P.S. I have been following Open Sim and Worldforge and so on for gosh it must be literally years by now. :)

I see what your saying.

Its like working but more like playing your mini movie from one perspective. And when someone joins the group he or she can enter the home or building of a library or company they enter and view mini movies from other perspectives. But it would all be real time or in an offline scenario. Or even make up a community of groups who what to help others make mini movies like what a real world movie production would do?

I can see the possibilities in this project if your going this route. If I am wrong please let me know.