I''m not sure how many people on here other than DavidRM have made any big multiplayer games, but I figured I''d ask anyway. What kind of development time did you have on Artifact (directed to DavidRM)? What kind of programming team was involved? Also, what kind of registrations have you had? I know you charge a one-time fee based on the ammount of time the user plans on spending, and I was thinking of doing that with a shooter I''m working on. Also, how did you go about designing the networking aspect, in what information do you relay from the server to each client? Only what specific units and such are around you, and only the land info when you need it? That seems like the most logical way. Thanks for the info, ~WarDekar

To answer your questions in stream-of-consciousness order...

Artifact design work began in August of 1996. We began looking for an artist in November, 1996, and began initial development in January, 1997. We finally found an artist in Summer, 1997, but work was slow because I bought a house and Dug had a kid, so it was late 1997 before significant development work was accomplished. We were both working full-time jobs through all of this, and that slowed us down, as well. In April, 1998, we started "closed beta" testing. It was March, 1999, when we opened the beta to the public and began charging. Official release date was October, 1999.

3 years of part-time development by 2 programmers wearing multiple "hats", 1 artist and 1 sound effects/music guy. Dug handled game design and sever-side programming. I coordinated the project, provided funding for software tools and servers to run on, did the client-side programming, and was just generally useful.

Right now we''re averaging 60-70 players online 24x7, with peaks of 120-130 players online at one time. We expect to grow significantly as school lets out in the US and summer vacation begins. Registrations are currently averaging about 2-3 per day, which will also go up as summer wanders in.

Summer, BTW, is The Best Time of the Year for online games. Christmas is nice, but nothing beats a good, long stretch of time-off for students.

As for how we designed the networking aspects...that''s a long story. Dug''s experience includes a *lot* of MUD coding from his college days. He did that for years, even after college. Our first game, Paintball NET, was based on what he learned from coding a MUD server. When we added the graphical client, we learned a lot about what was needed for that, and that "education" was applied to Artifact. And, of course, Artifact has taught us a lot that will be applied to our next project...

Opine Paranthetical: It''s a mild bummer that you can never really go back and properly "fix" the game or product that taught you what you know now. It''s just not cost effective, generally.

For Artifact, the map is divided into equal-sized segments. Unless you have units that are (a) in that segment or (b) positioned to *see* in that segment, you are not sent any information about what''s in that segment. However, we do send the *entire* map to the player. Since the map is static, sending it once works well.

I hope that helps and doesn''t repeat too much of what I''ve posted previously...


DavidRM
Samu Games

Thanks for the info. I think you''ve answered this before, maybe even from me, but what kind of costs are involved with the server and how many players/games are on each? Thanks again.

~WarDekar