So the biggest task in creating FOL (aside from basically learning to program from the ground up decently) is being efficient.

I may have tons of ideas and inspiration to make them look pretty, but I figure when it all boils down to it... If the game is laggy and hard to actually play it's a fruitless effort.

The reason why things may appear to progress slow (for those other skeptical game designers out there, who think games should be ready in a week or two) is because once I create a new section of programming (no matter how trivial) I spend as long as it takes on that section until I am sure it is as efficient as I can make it.

Today I did a fun little test to check myself. I created a 250x250 world with 5 Z levels and tried out creating a mass of duplicate mobs with a simple AI (much like the finalized one). I figured this would give me the best ideas on how the game will play because every zone will be 250x250 and then will be linked based on the map. So each world should have roughly 8-10 Z levels, none will have activity unless it is occupied.

Given I don't intend to run on my PC longer than necessary because it has a single core processor, but the test worked quite well.

Anyway firstly I created 1,000 duplicates of myself which all had a working basic AI (now note, NPCs won't have as many variables, especially lists, as these creations so it should be still more efficient)


It ran smoothly, it didn't even seem like I added any NPCs. I ran around pressing buttons and doing everything I could so far to generate any form of CPU usage. The only things different between now and when it is playable are a handful more graphics for overlays and icon_states changing. The only objects created after initial startup would be projectiles.

I figured this ran way too smooth for 1,000 AI NPC copies of me, so I tried a higher amount. Now note I doubt that there will be this many NPCs active at a given time, but I was still relatively impressed with the result. BYOND has alot more power than I heard.



According to the world the CPU went up 10%, but I didn't feel anything... The clients connected to me might but that's for a later testing day. And please note I had firefox and photoshop running while I did this. But as for gameplay I experienced no server-side lag in game. I ran around like an idiot pressing as many buttons as I could during their creation process and a while after they were created by verb.

For security purposes I created a slow loop in World/New() that every 30 seconds updates on the status, I figured if it does get out of hand for one reason or another (which it shouldn't, there are very few open loops) if CPU is higher than 50, which I feel is reasonable, tick lag will increase until the world can catch up with itself... If after a certain point it cannot catch up it will give a warning and reboot itself.
I've done everything I could to make sure the world shouldn't get out of control however.

I think programming-wise, this game's gonna be ready to play by first open test... But that's just hope, I would still like to try with a handful of people playing, I might even have a session later on when I host this for a few minutes and ask everyone to help me by constantly logging in and out, that'll give me an idea of what to expect if it has a big crowd upon testing, and maybe some ideas of how to handle it.

Well now that that's handled back to drawing *grumble grumble grumble*

<3 Ken