I'm using the recent ioq3 build, which I think is 1.36, on Windows Vista, with hunkmegs 192. 

I played with ioq3 1.36 linux-i386 Apr 12 2009 and encountered no problems, I think.

spy: dpmod with some model half-assed conversions from the Shambler's Castle Doom3 mod. It was just for fun. Does not work at all since there is only the one SP spawnpoint. ;) 

Great game too :-) Had to quit last night with 8/10 found. 

@metlslime, I am not sure why you are really negative about this map. Q3 has no scripting language, no logic entity types, no variables, no dialogue system, limited sound abilities (no volume, cannot turn off ambient sounds) and only two state moving objects with no toggle feature! What sort of puzzles were you expecting?

The first puzzle to be built was the generator with the outside decking. This was more a test of entity chains and trying to create a complex scripted sequence. It featured button mashing and each stage involved a greater amount of buttons than before. I stopped at three because most people get bored with anymore and wanted to move on. I thought it was interesting visually and with all the custom sound made sense as each stage changed the appearance of the machine. It is not a logic puzzle but a scripted event.

The second puzzle to be built was the chains and later on the metal block with steps around the side. Vanilla Q3 can only work with two state devices. A door is either open or closed and that really shits on the chance of complex puzzles. So I decided to prototype a lock with 4 tumblers to see how much fun it could be to try and unlock. Eventually I found that if each switch affects 3 parts at once then the puzzle could be a little more interesting.

I tested this puzzle on a couple of friends and it nicely highlighted the problem, it was easier to run around the room randomly hitting stuff until something happens. Not one friend hit switches and watched what happened, they prefered the headless chicken approach. I asked everyone why and they said 'It was not instantly recognisable what to do.' This puzzle rests on strong visuals and probably could of been designed better but I still think that people just don't have the patience anymore.

With Puzzle 3 and 5 being completed by random chance I decided to create what is puzzle 2 (you call puzzle 1) and make something that is less about random button mashing. So I decided to create a battery you charge up to open a door. I thought that a dead battery being charged up would be interesting visually as well as cool to highlight the door being locked. So I setup an entity chain that stores the battery state and forces the player to keep the battery level charged with switches. Initially had 4 levels but it was too complex so decided on 3 so people won't get bored constantly going back and forth betweem switches.

All the people that I tested the new puzzle on could not complete it. No one could understand the idea of charging something up via switches and just wanted to run around randomly mashing buttons. So I added loads of visual clues, cables merging into one, cables to the doors, lights on the switches, visuals on the battery etc and eventually some people got it. Even now it is the only puzzle that most people get stuck on and it makes me wonder about fps puzzles. I don't think anyone has the patience for them anymore.

Eventually I made a dumbed down version of this for puzzle 1 which does not drain the charge from the battery so everyone can just complete it with random button mashing. This puzzle does have a timing element hinted by the coils above, but hardly anyone understand it. I think this puzzle highlights that people in general don't look up!

I will admit having puzzle 1 and 2 next to each other is a design fault because most people think they are linked. I even had testers trying to run from room to room to try and solve them together! That is why the generator in the big window room is a gated area so that people did not wander too far to find stuff.

I also setup all the switches to have lights on them so people could understand when they were active but this seems to be ignored because I still see people hitting the switch. Some switches I actually lock in down positions but due to entity limitations could not do it for everywhere.

The center print text is all the engine offers for dialogue or feedback. You are right about the text content being awkward but the text developed over a long period of time.

The skulls are designed to be hard to find, if they were easy to find the map would be completed in 10 minutes. That means if you hate exploring or trying to find stuff then yes, you will hate playing this map. I did try to make things easier with the buttons and skulls all being the same style. The design fault is that I did not hint to the player that they are suppose to shoot them.

I am sure I do sound angry about my post above but I cannot understand how anyone expects complex puzzles from vanilla Q3. I thought what I did was impressive considering what there is to work with. I certainly don't expect everyone to like what I do, but I was assuming that people would judge it based on what the vanilla Q3 engine can offer. 

to working out how much patience I have :)

Probably ownt get to it for a few days though. 

My only complaints were about the puzzles, and I tried to give a honest account of a player's experience solving them for the first time, coupled with a designer's attempt to understand what happened and what could be improved.

It's really hard do do something this unique, especially as a mod because you are dealing with players that don't have prior knowledge about the game mechanics you're using, plus they are bringing in expectations from the original game (which is a different genre.)

But hopefully (if you make a sequel) you can get something constructive out of my feedback and don't think i'm just dumping on your level for the fun of it. 

maybe you misinterpreted, but the "bad:" section only contained those two bullet points, not the entire rest of the post. 

@metlslime, yeah I read you response as a couple of good things and then the rest as bad. I certainly don't think you are dumping on my map for the fun of it, I just did not expect what you said within the context of the Q3 engine.

I totall agree that people bring their gaming experience baggage with them while playing any map. I saw people (in the demos) randomly shooting stuff after they worked out what the buttons look like. Hardly anyone was looking up or observing what is going on around them. Some people were running around at break neck speeds that even I could not see what they were looking at or doing.

But I do need to stress that this map is not a MOD, it is just a map that can be run with the vanilla Q3 engine. This is the root of the problem when it comes to how complex and interesting the puzzles are. This map relies on what the default engine offers and does not create any new entity types or engine features. I am simply using what the engine offers and trying to see if anything interesting can be made from it.

In hindsight trying to twist the Q3 engine to create puzzles was probably a bad idea. The engine is just a skeleton of functionality that on countless occasion squashed my ideas for machinery. I wanted machines that could rotate, twist, open, build, develop, have GUIs and proper player interaction. None of this is going to happen, which why there will not be anymore to this series. It is just a once only experiment that taught me a lot of interesting points about puzzles.

* Always try to contain puzzles to a small area so people don't spend alot of time running around trying to figure what is going on. When people get frustrated they wander off and try to find if something else is missing.

* Use prefabs for all working components and teach the player about them gradually. For example the two switches in the first room and the door was locked so that the player finds them, recognises them and if they see them again knows what to do.

* Always put interactive information at player height because no one is going to look up or down. I really hoped that game minded people would be different, but everyone is the same. Nobody stares at the top of a building while trying to find the front door! :P

* Try to hint at what is going on with strong visual colours that stand out from their surroundings. Originally I had the working parts of the machines painted in subdued colours, but no one knew what to look at and frustration set in really quickly. 

http://www.4shared.com/file/237525207/3108c2d7/RSmith_GDC09_PuzzlesAsUI_ForWe.html 

What technical problems were you running into? I thought the whole point of q3a was to show that 'scripting' was basically turing complete?

Re: players don't look up or try to understand the puzzles... watch my demo. :)

I had a hard time with puzzle 1, because the 'active electricity' effect was too subtle for me to suggest different states.

too bad there's the 1024 entities limit, or it would be awesome to write a simple scripting language -> logic gates compiler 

cool presentation. Very interesting to read 

It's nothing new, but it fills in all the gaps. 

sock, have you considered taking this map into DP to get access to q1 progs entities? 

Editor : http://www.simonoc.com/images/design/maps_q3/moteof_source3l.jpg
Web : http://www.simonoc.com/pages/design/maps_q3/moteof_source.htm

The map source file is very large (10Mb) because I have included 200+ ASE files with relevant map files and a few of those are linked together to form other files. The ASE naming convention is a bit strange but the level was developered over 8 months with 156 map versions so some stuff will probably be in the wrong place. I recommend you turn on camera clipping in the editor, otherwise you will not be able to move the camera around.

The map is compressed into a small space in the editor so that the light grid has a better density on surfaces. I have included map compile options in the readme file and it works fine with q3map2 version 2.16. All the Q3A logic stuff is high above the map and is colour coded via textures to show the different functions used. Once I get a free moment I will sit down and go through some of the logic stuff because unless you are into scripting or coding it will be very difficult to follow. 

@necros, just use the source files I have just released.

@ijed, that is a nice presentation, thanks for the link. :) I liked the examples he showed for puzzles (portal/thief3).

@megaman, Q3 is just a skeleton of functions and it is really hard to do anything but DM stuff with it. You got 1024 entity limit across the map, 255 brush entity limit (doors, buttons). Two state movers that cannot toggle and rotating objects that don't stop moving and draw regardless of portals. A sound system with no volume control, no channel support or priority on sounds playing. No message HUD/text system besides center print. There is a good start! :p 

Would it be theoretically possible to spawn a bot in a cage outside the playing field and push him around with doors and platforms (possibly clones of existing machines) in order to create additional puzzle setups? For example, I could imagine a dial-like puzzle - you know, like those games where you have to turn several dials in opposite directions to move a ball or chip around. Something along these lines. 

i don't really know anything about q3 mapping or dp, i only mentioned it to see if you were interested. :S

the puzzles were a mixed bag for me. i understand that q3 has an extremely limited selection of scripting entities but it was still disappointing at times that some puzzles were easier to solve by just randomly hitting buttons.

also, just thought i'd mention this map looks absolutely fantastic. :) 

Was already in the Sock "fan club" but I like what I've read about all the comments on the idea of this map. 

That slideshow sums up a lot of what i would suggest to improve the puzzle design. Given the technical limitations of Q3 the puzzle mechanisms may not be able to change, so the best avenues of improvement are visual -- call attention to important things, limit distractions, improve feedback, emphasize state, represent identical function with consistent visual design, etc. -- and layout -- restricting player movement to focus their attention on a single room/puzzle at a time.

(many of these principles are already present in the level, e.g. consistent use of lights on switches, consistent use of the portcullis-needs-two-triggers mechanism, the only question is could they be pushed even further.) 

They seemed pretty good to me. The one that arrqf gur gvzvat gb punetr gur onggrel hc threw me for a while, as V jnf ybbxvat sbe n guveq fjvgpu naq qvqa'g guvax gb gel gur svefg fjvgpu ntnva... so that took me a few minutes. It was during this time that I noticed that gur fjvgpurf yvtug hc jura gurl orpbzr npgvir, which would've helped with the one that causes it to crash for me. It might just be my computer that has issues with that area. Strange...

I don't know why... but I expected it to be longer. I would've liked something that was twice the length, or more. But given how this place reacts to puzzles (this map and 768_negke), it was some way beyond your audience...

I look forward to the sequel. Even if no-one else here does... 

Loved your work, Sock. I'm not a MP guy, subsequently not a q3a guy, so just the concept of SP q3a is a winner, by me.

The Looks are absolutely gorgeous, as you know, and puzzles, well, apporachable, I think. Beside that, you provided LOTS more hints than anyone else here does on their own maps as how to beat the sucker.

Great job. 

and not for the sweet visuals alone ;) I enjoyed the puzzles, too -- I especially liked how the triggers/switches went inactive after they were completed -- but a few of the skull and target placements were a bit tough to find.

Great job all around, Sock! 

Nice way of doing spoilers on this board! Good job!

...course, if we could get some kind of spoiler tag, that would be even better (and save the few seconds it takes to translate). Hint. Nudge.

It could be as simple as automating a [SPOILER] prompt before the text in said tags, and then setting the text in question to the background color to force the reader to highlight...

Also, I don't know why the morons hate the puzzles either. 

There's even a Firefox plugin that will translate it for you... 

it would have been nice if you made the whole paragraph rot13 though...

http://operawiki.info/BookMarklets has a nifty bookmarklet for it. :)

pjw:



! :-)