View this PageEdit this PageAttachments to this PageHistory of this PageHomeRecent ChangesSearch the SwikiHelp Guide
Hotspots: Admin Pages | Turn-in Site |
Current Links: Cases Final Project Summer 2007

Sp02 Final Exam Review: Design Networked Games

A) Absolutely. Virtual machines are great for ensuring cross-platform compatibility.

B) It would depend on whether the game is supposed to be 2D or 3D.

C) MIDI would probably be a great choice for background music, because of it's small size and the fact that there's probably going to be a lot of it. Also, playing MIDIs generally requires a lot less from the CPU. However, MIDIs can't really handle voices, so MP3 files (or some equivalent thereof) would probably be necessary for the voices.

On (B), is there a way to do 3D without sending the whole model? Mark Guzdial

Sure, you could just send a texture of some sort and paste it on a plane, but to make it look right, you'd really need to have multiple images of each frame, with the characters pointing in various directions. With a 3D model, you don't have that problem, so I'd vote for the 3D model. - Mitch Halpin

As we discussed in class, there is a reasonable way of doing 3D without the whole model. Think the original Doom. Mark Guzdial

I tried to look up stuff on the Original Doom. It talked about something called pixmap blitting. Unfortunately that's where I got lost. The only alternative I know to these 3-D models is the flash substitute. Could someone discuss the original Doom implementation? Jai Kejriwal

What's the general issue, in sampled sound vs. MIDI, in flash vs. MPEG? Mark Guzdial

MIDI and flash deal with abstractions of data that tend to be smaller in size and less precise, maybe easier to manipulate.
MPEG and sampled sound are very precise descriptions that require more space to store.
Rick Giuly

A) I agree with Mitch, one of the main plus points of VMs is that they are cross platform compatable, so that's absolutely the way to go there.

B) Another way to do this to make the models appear 3D but without the overhead of the actual 3D model (and I believe what they did in Doom) is to take snapshots (2D images) of the characters in multiple positions and at multiple distances then simply change the image when the character turns or gets closer or farther away. You can even make a 3D model then use that to take your snapshots, so that way you get reasonable looking characters but you don't have to send the whole model accross the network.

C)I wouldn't go with midi for this because it would be impossible to create the correct sounding theme music and the voices would sound terrible. Kids will want to hear the voices that they hear on TV, so you'll have to go with a compressed format like wav or mp3. I'd probably go with mp3 here because you need a relatively small file for transfer, but it will also give you cd quality playback, which is what you really need here.

Bill Branan

A) Yeah, VM's fine, as long as it has libraries available to support the things you want to do well (don't want to redo that).

B) This depends. Personally, I'd think if there were a fixed number of characters, the ideal thing to do is to store all the 3D models locally, and then just send character and animation instructions across the network. Doom's solution looks terrible and is outdated, necessary only for the lack of 3D hardware at the time, and I wouldn't recommend it.

C) Again, all this can be stored locally, so I recommend high fidelity. Modern systems can handle this. MP3's for everything.

And for the above person... WAVs are not compressed! They have some primitive constructs for repeating certain blocks, but if you play a WAV file as a raw sound file you won't be able to hear the difference 99.9% of the time. I wouldn't recommend WAVs; the extra quality is inaudible for the ridiculous file size increase.

So here's a consensus opinion, of sorts, with my two bits:

A) VM? "Uh, yeah, sure..." I think we've covered fairly well that VM==cross platform; say otherwise and you will be replaced with a very small shell script...

B) About 3D versus alternatives. How about (1) storing/rendering models locally, and then (2) when the characters are moving around, all we need to distribute to the clients are simple descriptions of where everyone is. For example "Angelica moved 50 left, her left arm is up, and her face is now... blue." Flash also works, shrinking and growing as we move back and forth and using different discrete angles to provide an illusion of 3D. In fact, the silly snapping when characters turned would go right along with the Rugrats look and feel.... aww, come on, didn't you folks watch them when _you_ were little???

C) We beat this one into the ground a little as well: high fidelity versus sampled, compressed formats. MIDI for background music; most commercial instrumental (and Britney-type) music nowadays ends up being MIDI on steroids anyway, and low-sampled/compressed .wav or MP3 for voices.

alien song

Link to this Page