| |
Demo Alarm: Time Keepers Displaying 1-10 of 10 total.
1
Gayo
|
You might assume, since the front page has been quiet, that not much is going on. It's actually been pretty active, though! I've just been very busy with my meth habit and haven't had time to post. So, what's new? First up: Ness has released a second demo of [Time Keepers] (offsite link).
You might remember Time Keepers from the original Demo Alarm, way back in 2005. This is basically the same game, but with some added functionality, a second and much more complex mission, and a bit more actual plot. For those of you just tuning in, this is an "infiltrate and steal stuff, optionally killing everybody" type game. Most of the original problems (inadequate weapon choice, slow text speed, no saves) have been fixed -- the demo now autosaves regularly as well as manual saves, and you have a wide selection of weapons to choose from. THe button selection is a bit complex, but the game takes beautifully to a gamepad. The enemies are still pretty stupid, so gaming the AI is a highly effective strategy, but on the whole it feels a lot more like a game and less like a proof of concept.
 There are Achievements. What the hell, Ness! I think these are the only two currently in the game, but still, wow. It's nice to see someone making good use of the net features this way; this definitely raises the bar for other VERGE games. Therre's a lot of potential for more of these as the game progresses -- the existing missions are heavily constrained, but it would be easy to design more open missions using this same framework, and even now you have a lot of options in taking out the guards.
 Time Keepers still feels very alpha. There are some pretty big bugs, and the game desperately needs a proofreader. That said, the main problem is just that the AI still has a long way to go: you can just stand off to the side and a guard will lose track of you, for instance. The second mission takes place in a place where armed guards swarm the instant you destroy a security camera, but you can break into a guy's room and loot in in front of him and he doesn't react. In a sense this is a necessity since outwitting the computer is the main way you overcome the enemy's superior forces, but there's definitely some more polish needed.
What flaws Time Keepers has can be forgiven because it's an extremely ambitious game. I'm really glad to see this sort of neat stuff being done with VERGE, and it's nice to see 2D tactical infiltration games more complex than "move when the guards aren't nearby," especially fanmade ones. Whereas the original demo was more promising than truly fun, this one is actually a lot of fun to play, and I look forward to seeing what the next mission will be.
Posted on 2008-06-29 19:46:19 (last edited on 2008-06-29 19:48:51)
|
Biggs
|
Hi,
Cool post & stuff! TK is one of my favorites!
On a different note: From what I had learned about netcode in the past, I had gotten the impression that it's functionality was either limited or exremely inefficient, thus rendering it unsuitable for realtime network gaming. Is this still the case? How much realtime, network-intense stuff can Verge handle?
Posted on 2008-06-30 16:46:26
|
Gayo
|
That's a question I would be uncomfortable answering without practical tests under my belt, and I don't think anyone has done that. Ustor's turn-based network game worked plenty nice, and I think VERGE just uses sockets so it shouldn't be substantially slower than any other socket-based message-passing game.
The main idiosyncrasies of VERGE's netcode, as best I recall, are: it uses a hardcoded port, and it both gives and expects a VERGE-specific header for every communication. This may have been changed, as there are tons of undocumented changes since the last public build, but I think it's still the case.
That said, there are some major problems you'll run into in trying to do a very dynamic game with netcode. Most of these are the same problems you'd run into anywhere -- the game needs to predict the movements of other players to minimize the effects of latency, and decisions like which party's client does the important calculations will have a major effect on what bugs you run into and what hacks and exploits are possible.
Also, VERGE as a whole is inefficient and hogs cycles. This isn't a problem for most clients, and you should be fine for games where one party hosts and others connect to him all using copies of the same software, but if you plan to have a centralized server, whether to manage play, to arrange groups, or simply to collect statistics as with TK, you may be better off finding out the format for VERGE's message headers and then designing the server in a "real" language.
Posted on 2008-07-01 01:16:31 (last edited on 2008-07-01 01:18:00)
|
mcgrue
|
Yeah, Gayo's got it correct as of present: VERGE has no concept of asynchronous events right now. Your server loop is done via polling.
This isn't so terribly bad since _EVERYTHING_ in verge is basically accomplished via polling, but it's still a bit on the amatuerish side. Even so, it's a fully capable amaturish side.
I'd like to move a lot of things in verge away from that eventually, going to a far, far, far more event-driven model, but that's in the future and not in the now. Any polling model will still work in the perfect event-driven future, though, so feel free to continue working on any projects using it here in the present.
Posted on 2008-07-01 02:18:07
|
Ness
|
Weee!
One of the best things about these write ups is that people point out the biggest flaws, more so then random play testers. I've put improving the pathing / AI at the top of my list, and fixed up a large chunk of the pathing last night. There's still some AI work to be done.
I can neither confirm nor deny the existence of up to 4 achievements with more coming in later releases.
In terms of the netcode, I believe the latest SVN revision has the ability to open an socket as well as send and receive raw text. I haven't used them yet since it's a bit too much for what I need.
Posted on 2008-07-01 09:28:37
|
Biggs
|
Ok, cool; thanks for the good feedback!
So far it sounds like V3 netcode will do good. Polling is fine with me, since I like using coroutines with Lua, which is cooperative multithreading anyway.
So I suppose, then, that a V3 client can be simultaneously connected to multiple hosts at a time, and can be both a client and a server (peer-2-peer-style).
Regards
Posted on 2008-07-01 14:47:36 (last edited on 2008-07-01 14:49:59)
|
mcgrue
|
As far as I know, yes.
I'd probably recommend a single server, many clients model unless you're experienced with netcode.
Posted on 2008-07-01 15:27:42
|
Gayo
|
Ah, Ness is here! I didn't want to talk too much about the bugs and AI issues since Demo Alarms are mostly puff, and I did like the game, but since you're interested I'll be happy to discuss it a bit more.
As for bugs, I forget most of them, but the one that stuck out was that the training segment interacts very weirdly with the rest of the game -- you can skip it and do it later, or leave in the middle and pick it up later, and I think I might once have done it twice, but I'm not positive.
Mostly it was the AI stuff that was an issue. There's a lot that could be added to AI -- have guards notice whether doors are open or closed, have them explore a bit more in alert mode, and so on -- but the main stuff right now is just their perceptive ability. I had this one showdown in the casino security room where the guard was chasing me around a desk over and over again, Looney-Tunes-style, and eventually I got him trapped on the opposite side of the desk, couldn't see me, and gave up. What I think you want to do here is have two kinds of obstructions (coded by tile): those that stop movement only, and those that stop movement, vision, and fire. You could even make it really complex, and divide the former type into "low" obstructions that can be ducked behind to also obscure vision and fire, and "uncrossable floor" obstructions like pits that can't. As for other bits of weirdness, the sight cone seems a bit strange -- I was able to stand a couple tiles to a guard's left and walk behind him without alerting him, and there were a couple situations where someone I was right next to lost track of me.
Whoever said that the targetting is a bit onerous was right -- honestly I think that a game like this would work best with free-directional mouse-based movement and firing, but this is pretty good, it just could stand to be simplified a bit.
Posted on 2008-07-02 23:07:16
|
Syn
|
Wow, this game rocks. I am actually having fun playing it and not just testing it. Let me provide you with useful comments.
The enemy spotting you before he appears on the screen is not fun. The thing is that there should be a way to know where the enemy is facing most of the time. Either on the mini-map(MGS style) or making sure they can't spot you if they aren't at least partly on screen.
Some people suggest a different aiming system, I personally don't think this one is that bad, although there should be a way to make sure that you can control who you want to target more easily. One thing that seems useless is targeting items before pickup, it seems like a useless extra step when you could have it work simply by pressing space since it doesn't do much on it's own.
The thing I hated most about this game was the eye hurting font. Not only is it weird it is way too small. I really had my eyes hurting on me after a while.
The AI needs a tweak, but it didn't annoy me as much.
Some of the things I love in this game is that it wasn't medieval. (No elves, woo hoo!) It felt really like you were exploring the maps and not just wandering in it. Had nice overall feel and nice extras like achievements, scoreboard, locker hidding, lock picking, etc. Not only that, I feel the system you set up is a good foundation for adding many original items and mechanics later.
eg: A sword that can work like a knife even when in front of enemy. A item that makes noise. Learning to pick locks faster. Dressing up in different clothes much like the in Hitman. Setting up traps. Sniping. A crouching button. Using ventilation shafts. Etc.
If you are interested in some help or making a team, check out my post here:
http://www.dreamstem.com/?p=17
Oh, and btw I totally killed those time police guy ambushing me at the end and I still lost!
Posted on 2008-08-18 23:47:18
|
Ness
|
:D
I'm still tweaking the enemy sight range's and angles. I made a decision a while back to not show the enemy's FoV on the radar. Although 3-4 years ago I was also planning more player punishment in the game so...we'll see!
There is a targeting system in place right now that works on priority. So it will pick a enemy with a shotgun over an enemy with a pistol, and a enemy with a pistol over a civilian and so on. I started work on a mouse control system in the hopes of being able to use Gamepad, Keyboard and Mouse, but mouse isn't cooperating right now.
The decision to use a 4*4 font was mainly because I wanted to get text that isn't completely necessary to the game in there without taking up a lot of space at 320*240. Also I usually play at full screen so it doesn't bother me as much...or maybe I'm just use to it.
You actually named more then one thing I'm planning for a future release :)
Posted on 2008-08-19 13:14:22
|
Displaying 1-10 of 10 total.
1
|
|
