My "Contingency Plan" design sketch

[greendragoon]

(Remember, as recently as Myst 4 we were clicking from room to room in discrete steps. Would the feel be so different if you saw other players doing that?) I'll just send movement packets one-tenth as often. Or one-twentieth if I have to.

Yeah, but we didn't have to look at anyone else in Myst 4. If this ends up looking worse than Uru, what's the point? why don't we just beg Cyan for a version of Until Uru.

That aside, there is also the question of future expansion. I have heard many say before that one of the things holding Uru back is the aging engine. I would hate to make the same mistake.

(The basic idea of jumping from server to server is simple. However, then you have to grapple with other server-synchronization problems: Identifying "the same player" between servers. Keeping your avatar clothing the same. Sending KI messages between servers. These are solvable problems; I just don't have a solution ready.)

That's why I see having a central server is essential. Each player could have a unique identifier. But there would still need to be a central server to keep the master list. KI Mail could also be routed the same way. As for clothing. The client could keep track of the clothing and just let whatever server it connects to know about it.

[Lontahv]

Oh dear... NO way are packets going to be sent "one-tenth as often". That makes the bez-curve based interpolation even harder.

And THIS IS A 3D GAME, this is not a 2d game like M4. And this is as much an exploring-game as it is a adventure game. People don't want other people to look like their falling off a cliff when their not.

Even if the packets were sent every 8 seconds, that still doesn't mean we can be stupid about the protocol, we may have people on DIAL-UP, we can't have bloat in the transfer. What about this: You do the chat and emotes, and I do the avatar movement protocol and bezier-avatar-interpolation, I really think that hex is better than ascii when it comes to large amounts of data being transfered--Paradox can you back me up on this?--I think that linking and lots of ages is overrated, if we have ages each one should be special--like Cyan wanted Uru to be in the beginning.

Imagine having 100people in one place and not being laggy, we need this protocol to be GOOD we can't have some kind of ascii stuff dictating what we can and can't have when it comes to crowds--if you don't believe me go and talk to the Uru, SecondLife, and WorldOfWarcraft devs. At first I was scared of hex but... I've grown used to it and think that it's a much better transfer than ascii. I think that your just uncomfortable with using hex because it's strange. Get a hex editor, type some text in a txt file, and then open it in the hex editor, you'll see TONS of data.

I really don't mean to be controlling or mean--I just want this to be done right. If this protocol is bad and this game becomes popular... who's going to say: "hey let's upgrade the whole protocol", what's done is done. Think of what you could use the extra bandwith you get by not using ascii.

If you use ascii commands and tags then...

A) You'd need a really fast server connection

B) I might not be able to play this game because my upload-speed is puny.

Waste is NEVER a good thing, especially on the internet.


~Lontahv

[Lontahv]

This is what needs to be sent for an emote with Jabber:

3c 6d 65 73 73 61 67 65 20 74 6f 3d 27 65 64 65 72 66 6f 6f 32 37 40 75 72 75 2e 6e 65 74 27 20 66 72 6f 6d 3d 27 65 64 65 72 66 6f 6f 32 37 40 75 72 75 2e 6e 65 74 2f 62 65 6c 66 6f 72 64 27 20 74 79 70 65 3d 27 67 72 6f 75 70 63 68 61 74 27 3e a 3c 62 6f 64 79 3e 62 65 6c 66 6f 72 64 20 77 61 76 65 73 20 68 65 6c 6c 6f 2e 3c 2f 62 6f 64 79 3e a 3c 65 6d 6f 74 65 20 78 6d 6c 6e 73 3d 27 75 72 6e 3a 75 72 75 27 3e 77 61 76 65 3c 2f 65 6d 6f 74 65 3e a 3c 2f 6d 65 73 73 61 67 65 3e

Here's what it would look like in my system:

<prefix(1/2byte)><DefinedPlayerID(1byte)><emoteID(1byte)>

it would look something like this:

1c 00 01 8c a8

See the difference?


~Lontahv


EDIT: I'm not asking you to use hex, I'm only asking you to let me use hex.

[Lontahv]

And my plan is that every time you log in you download the protocol again. In a list like this:

End = cc
Begin = cb
Emote_wave(chat trigger: $player waves) = 8c a8
etc.

This way the protocol could be updated without the client having to be rebuilt.
It would mean that vast server-changes and new emotes could be added without too much to-do.
And then it knows how to send AND receive data with the new protocol.

~Lontahv

Mod: sorry for triple-posting. *ouch*

[belford]

If this ends up looking worse than Uru, what's the point? why don't we just beg Cyan for a version of Until Uru.

Answering your question literally: because this is the case when there is no Until Uru, and Cyan has not responded to begging.

You don't have to explain to me the difference between the byte counts. I've explained the tradeoff: keep the environments looking perfect, and take the hit on the avatars.

(Someone watching me play MOUL remarked that Cyan has *already* made this tradeoff. He was astonished by the quality of the Neighborhood; then I turned 90 degrees, and he was astonished by (what he saw as) the crude avatars standing stiffly around *and in* the fountain, in a totally unrealistic way. I merely pass this along as an outsider's viewpoint.)

People don't want other people to look like their falling off a cliff when their not.

That much is easy to avoid. Really I'd rather have avatars standing still than using extrapolated motion. It looked bad with even a second or two of latency, which happened often enough in MOUL.

I'm fully aware that this falls below MOUL's technical level, in certain areas. It surpasses MOUL in others. I'd argue that MOUL took the approach of being hand-written, as fast as possible, and sloppy under stress -- and that proved *unworkable*.

But like I said -- my post is not a package deal. Mine any bits you want out of it.

[Lontahv]

The protocol and the physics and the "hopping" avatars are the only things that I quibble with, the rest is just as I'd do it. I'm not sure how much dreamer and how much builder you are but... I'd love to work on a team with you to build D'Nay. I don't think a lot of people will like the avatars not walking and rather hopping.

The thing is that Uru's very different from the other Myst games and I think people want D'Nay to be a "fan made second Uru", if you watch in the cavern--most of what happens if people run around. It's just part of the culture. And to make the avatars not walk but "hop" when they move... it strips some of the Uru-y goodness out of D'Nay. We aren't trying to make a multi-player riven here, we're trying to make a royalty-free/legal Uru copy that the displaced Uru fans can hangout in.


~Lontahv

[Nek'rahm]

Very good ideas here belford.

I mean, heck, we could have an MMO take place in another Cavern...

I dunno, have another Shaft be found in Europe instead of North America. Would be a wicked way to restart the whole storyline of Uru in a crazy way

[Lontahv]

Hey, what about if the Myst-fans go on spelunking expeditions in D'Nay?

~Lontahv

[Lade'e]

Hope you don't mind a non-programmer dropping by and adding her 2-cent's worth ...

As I was reading the plan and ensuing conversation, I kept thinking, "Why can't the fan-created game be a franchise of Cyan's?" I've been reading a lot of these threads, here and on other forums, and unless I've missed it, no one's suggested it yet. I don't mean like Wendy's or McDonald's, but more like StarGate/StarGate Atlantis or Star Trek/Star Trek Voyager or the CSI series. And the CSI as well as the Law & Order series are good examples of how they can be marketed (targeted) to different audiences - they're each on different networks, for instance.

Another thought had occurred to me as I was reading and that was that the fan-created game wouldn't necessarily have to happen on Earth. Surely when the Ronay left Garternay there must have been some dissident faction that wrote a new home somewhere else? (Seems highly likely to me.) Could Ae'gura possibly have handled all those who fled? (Seems unlikely to me.) However, I get that part of the fascination for the explorers is that it's also part of their history ... somehow. But, the advantage to being off-planet is that cultures evolve differently. Therefore references to the D'ni civilization we know and love can be morphed into something less likely to violate copyright and trademarks.

However, if kept on earth ... oh, my ... all the places left, yet, to explore. All those cliff-dwelling & cave-dwelling civilizations that are archaeological mysteries could be explored.

Anyway ... just some food for thought. Especially with the news that GT is launching the single-player version of URU on April 10th. I'm guessing, but I'm betting that GT is going to hold onto their copyrights and we won't see shards any time soon.

Lade'e

[Aloys]

That aside, there is also the question of future expansion. I have heard many say before that one of the things holding Uru back is the aging engine. I would hate to make the same mistake.

Assuming you are talking about rendering engine some assumptions needs to be put to rest here. (the network part of Uru could certainly be discussed too, but that's not my area of expertise)
While today there are certainly better engines than Uru's, using them is more difficult and a LOT more time consuming.. (that is if you want to use all the features). It would be great to see an Uru/Myst-style game on a modern engine, it could look absolutely awesome, but that's definitely not something a small fan team (even a very dedicated one) can do; unless you have very skilled people and want to spend 3 years on it.

Going a little technical here:

click Show Spoiler

Using a more advanced renderer supporting normal mapping, advanced shading system, soft shadows, HRD, etc.. is really cool but if you want to use those features you must spend a lot more time creating your assets. If you want to use normal maps then you must create high-poly models to generate them, if you want to use nice sharders you also need all kinds of various textures and that takes time too, etc etc. If you know a little bit about mods in FPS you might know that the number of mods created over the last couple years has dropped dramatically because of this. It just takes too much work to create decent art for modern engines now. The usual metric is that it takes at least 4 times more work to create art for a current game than for games of the previous generation.

Also, Uru's Plasma may be old by some standards, but it works well and allows good artists to create great stuff. When I look at Uru and compare it to some more 'advanced' gamed there's nothing to be ashamed of. Some features may be lacking and some could be improved, but for the most part it does the job well.