Whoah, awesome that you ported this to UDK. Looking forward to testing and posting something more constructive

Yeah, actually when I think about when I had lost my robot amongst the other stuff, I was running round in circles (maybe to try and pick out which one was mine) panicking as the next wave came in. Haven't really replicated the situation with the big highlight so not sure what would happen. This is why I suggested a different color. Howver if it's technically hard to change the colour, a permanent alpha circle on top of the robot might do the same thing? Not sure how it would look though.

BTW the badges are a really nice idea. Would be nice maybe to have a third "find the combined optimum" type badge, or just a highscore tracker or something to also encourage the "middle ground"? Just an idea...

I have played a bit with weapons-only: The weapons are very satisfying, especially the railgun, RPG and pipebomb. This was mostly on the axebot-only level, so there was a lot of furious backpedalling (to be expected I guess), but getting the following bots into position for a bomb blast or rail shot was good. Another random low-priority idea: If (some?) axebots had some spreading-out behaviour so they came in a line abreast it would add a bit of variety to that level, but no big deal.

Oh, and since I was playing weapons-only the "fire a bullet whilst clicking a menu button" cropped up quite a lot. If I can reliably reproduce it I'll let you know.

AND finally... would it be possible when buying weapons and armour to automatically set it as "selected" in the inventory panel and (perhaps) set it as the current weapon/armour?
It seems like if you just bought something, you intend to use it pretty soon, you don't want to be switching tab and clicking stuff in the menu when the next wave of enemies is coming.

Glad to hear you removed the forced delay on death, I felt that 5 seconds quite a few times...

"capitalist" being considered an insult in europe

errr what...?

Also, the sticky walls thing (and enemies failing to path around walls) becomes totally irrelevant (in a good way!) in the main game, because there are no walls.

I will get the latest build and have a got in arena-shmup style like I said I would...

Like I might have already said, I would actually be disappointed not to be able to spend $5-$10 on this, assuming some (not necessarily all) of the rough edges were smoothed off.
Only one copy though

I just discovered this and have been chuckling my way through the videos. Looking forward to seeing more!

Ah OK. I was thinking more for file/class scope stuff, i.e. replacing
#define MAX_HEADROOM 12
with
(static?) const int MAX_HEADROOM = 12

I guess my guidelines (C++ and C#) come from team projects that I've working on at various companies, and for the most part they have converged on the same thing, except for a few minor details like variable prefixes, indent style, etc.

Also from
http://www.amazon.co.uk/Effective-Specific-Addison-Wesley-Professional-Computing/dp/0201924889
and
http://www.parashift.com/c++-faq-lite/

I've always used an IDE which helps, for example autocomplete allows you to have longer descriptive variable names without much impact on typing speed and IDEs generally make it much easy to create and navigate between multiple source files, and don't have to create makefiles.

...NamesAreUsuallyCapitalised like that.
...I pretty much never reuse local variables for multiple things as it just gets confusing and invites bugs. Also might suggest the function is too long. Variables are declared where first used and tightly scoped.
..."static const" variables and inline functions instead of #defines where possible to avoid errors.
...I try to have functions and classes that do one clear, atomic-ish job.
...tons of formatting habits, like 4-space tabs, {} on new line, yadda yadda.
...etc, etc. Of course some of this goes out the window if really rushing to finish something.

I still shudder when I remember 2 months spent last year at work maintaining, refactoring and then finally just deleting a 5000 line function with ~80 goto statments, and variables that were all declared at the top and arbitrarily did ("id3") or didn't ("id4") retain values for future whilst the code skipped around the blocks of the function.

Heh oops, didn't see this thread - I just posted a massive braindump on a similar subject in another thread:

http://forums.tigsource.com/index.php?topic=11286.msg345918#msg345918

Not exactly the same topic, but there's a big overlap. I totally agree with Zaphos and Salt's points. Great tutorial/write-up Mr Salt 
I tend to get too attached to stuff that I know, like A*, boids, and don't experiment enough with other stuff like the heightmap system.

Regarding pathfinding, A* (or whatever you use to plot a route from A to B) is the easy part.
All these algorithms are designed for finding a route through a field a static obstacles (or more abstractly, a static node network).

If your units don't block pathfinding, for example if they are human soliders and the smallest gap between obstacles is large then your problems are greatly reduced. However this is fairly unlikely.

Imagine you implement A* and allow each tank to request paths individually from the pathing system.

If you send a single tank from A to B, no problem, it will look great.

In a more realistic but still simple example, you might have a group of 5 tanks at point A on one side of an open plain and you send them to point B at the other side. In this case, all the tanks will request paths (problem 1: high bursts of load on the PF system) from their current position to point B (problem 2: they can't all exactly get to B, they need to be arrayed sensibly around B) and they all set off. Problem 3 is that naively they will perceive each other as obstacles to route around. This leads to some stupid looking wiggly paths that are immediately even more stupid because the tank they were pathing around has also started moving, and has almost certainly moving into the path of another tank.

Problem 4 is that other unrelated units could be moving across the path, invalidating it to some degree. So maybe you should periodically repath? In fact, you definitely should, but you still need to solve Problem 3 as it will keep coming up each time you get a new path.

That's the simple case. If your map is more like a set of corridors (like a city) then corridors can be blocked by packs of units. Clearly if these units are static then A* will just give you the "next best" corridor if there is one. However, if they are moving then maybe they'll have gone by the time you get there.

Bonus problem: Units might have different abilities to cross rivers, crush obstacles, etc, so you have to design the navigation data to allow this.

So, yeah, have fun!  

I have implemented some AI to handle this in the past, but it was far from perfect, so I won't tell you what game it was  

IMO the main things are:
- if units are moving as a group, have some system to group them into squads/convoys for the duration of the movement. Ignore squad mates when pathing, only the "leader" requests a path, but make sure you consider the group when choosing a path (e.g. don't choose a route that some members can't do, or maybe split the group if needed)
- implement (or find a 3rd party implementation of) some reactive steering to follow the path and divert round minor obstacles and other units (see this)
- somehow give units a way to advertise to the pathing system what they are doing, so that other units can make a sensible choice about whether to go around them or their future position, or ignore them.
- design the underlying data with all these problems in mind, e.g. a tile/node on the grid knows what is currently blocking it

I haven't really studied recent RTSs for this stuff, but maybe I'll take a look at some later.
Hopefully it gives you some idea of the scope of the problem:
                                                             

(Oh, final bonus problem: If the unit is sent to attack something, you can just give the target unit's position as the destination of the path, but that will often end up looking stupid for long-range units in obstacle-heavy areas. You might want to do some clever analyse of good attack positions for a given target, or just make the player solve it with micromanagement)

edit: Useful links: http://www.cgf-ai.com/links.html

I had a quick go on this and it seems very promising, especially the desolate and desperate atmosphere.
I really think you need more transparency though, for example this:

As I gather, if you don't have boots, you can injure yourself when you sprint on rocky or rough terrain.

That is correct. I didn't want the character to be able to run at full speed through the entire map, so the rubble and debris is designed to force you to think a little bit, and slow down while you explore.

I like the idea, now that I've read about it, but as a player I had no idea what was going on. Could you add an "event log" section to the UI like in traditional roguelikes? So you'd see something like "You stub your toe on a rock and fall." This would also be useful in combat, although the difference between this sort of thing and combat is it would be much harder to represent this in sprites compared to "hit/miss/kill" in melee. And at least you can guess the spider/dog/things are dangerous.

Also, another vote for turning down the "dust storm" screen effect, or perhaps making it more of an occasional hazard than a constant thing. It hurts my eyes too much at the moment.

Anyway, keep it up, looking forward to seeing where this goes! 

Sounds like good work... just to clarify something regarding selling:

Yeah that happens, it's a bit of a prick but having the sell icon disappear after you sell something would get annoying if you wanted to sell multiple things at once,

I'm not suggesting it disappearing after you sell something, just being much more obvious you're in "sell" mode (cursor not going underneath things, text prompt at the bottom, whatever). Even bug-free modal UIs are dangerous beasts.

Haven't tried the armours yet - I might have another go at some earlier areas in arena-shooter mode and try them then.

OK, failing colour, I think something moving would be best, like a spinning ring or something. I died once just now because I was getting axed and couldn't see where I was...

I was playing around with Juggernauts after selling all my buildings at the end and found a couple of bugs:

First, after selling all my stuff, I clicked the button to buy a juggernaut. Then, when I placed it, I was still in "Sell" mode, so it immediately exploded when I placed it!

Second, after driving around the 2nd juggernaut, opening the front and letting out the axebots, I tried to sell it and got this crash:

Code:

___________________________________________
ERROR in
action number 1
of  Step Event
for object PlayerVehicle0307:

Error in code at line 22:
               if (distance_to_point(Target.x,Target.y) > 250)
                                            ^
at position 43: Unknown variable x

I tried the vehicles in general a bit more this time, and I'm even more sure there should be an enter/exit button. Also the way the vehicle always follows the cursor means it drives round in an out-of-control fashion if you're going to click something in the menus. Maybe hold right button to drive?

Ah ok, will download and check it out. I still have to beat the last level

Where's the 1.1 download link? All the links seem to point to "http://www.indiebird.com/download.php?f=AotP_Zombies_1_0.zip"

Totally agree with AuthenticKaizen. I would have been properly angry if the DVD drive started randomly opening as my PC is in a confined space and it might have broken something. You're saying you already took this out before the current version right?

As for plot: Tetris doesn't have a plot and no-one complains about that!

- The whole button thing with firing while buying stuff and not firing when you need to fire is fucked. I need to redo the whole system for that. Just the other day I was testing it and fired a rocket when I wasn't supposed to and got really pissed off.

- The whole menu is a bit of a mess, I hate making menus.

Rather than enable/disable the fire control when entering/leaving the menu, how about having the menu consume the mouse event if the pointer is inside the menu area and pass it on to the player controls if not? I think this is a fairly standard way of doing it.

- The player is hard to see, I gave it a faint glow which none of the other bots had, but I think it needs to be darker. Also, making it so the player is always on top of everything else would be good. Currently, they go underneath the tops of turrets and the arms on the factories. I think it was to give it a sense of depth, but in reality, it interrupts the gameplay too much. Originally, I had it so you couldn't walk through buildings (like real life) but it sucked so hard I had to take it out.

I'm not at all convinced either of those things would make the player much easier to see. You need something bolder, like a different colour, or an arrow. The problem is that if the player has been doing other stuff (I'm not sure what I was doing, maybe even just watching the moneymite picking up coins), and then the next wave starts, they need to *instantly* see where they are. Seriously, this was a major frustration.

- The reason the vehicles work the way they do is because I wanted to try something different

Fair enough. I hadn't tried the armours yet as they seemed expensive and I was happy just building stuff. I hope the anger at dying the first time I try the juggernaut doesn't put me off the game... 

- I did make an endless level...

Yeah, I see your points here. That 10-level area + highscore sounds like a good alternative.
If you hit technical limits, how about increasing health of the enemies as the game progresses, rather than adding more?

I'm also happy you enjoyed most of it

I made it sound like I enjoyed it less that I did - I could happily play it a lot more, although some of the bugs might eventually drive me away. As it is, I'll wait for a little while and hope for a patch!

Hi, I just discovered this and have played 3 (or 4?) areas. Great little game!

Here's my feedback (bad stuff first, because even though I really enjoyed it, I'm a negative critical bastard)

* By far the major problem was not being able to see myself among the throng of other robots. I was very close to ragequitting at one point because of this. Perhaps you could tint the player pale green or red?
* I really felt the 4th-wall-breaking stuff didn't work in this context. The "error" boxes are far too much like real errors. My girlfriend sitting next to me was saying "wow, there's a lot of errors in this game!" Personally I just found them annoying and clashed with the rest of the game... I was happily immersed in a neatly constructed RTSish thing, and then had this glitchy stuff flashing up. I acquired an instinctive "oh fuck off!" response after a while.
* Quite often when clicking on a button in the UI, my guy would also fire. I can imagine this being really annoying if ammo was tight or you had an expensive gun equipped.
* Some of the prompts in the tutorial (which was otherwise very nice) were out of order, i.e. you had to break down a down to get to the message that said "click to attack".
* I'd agree about the UI layout - if you can improve the layout of the tabs (all on one page, popups for info?) that would be good. I didn't find it too bad though.
* The dark corners of the playing area sometime made it possible for cannonbots to lurk out there and be very hard to deal with
* The minimal style is great, never have medium-sized circles felt so menacing
* The AI behaviours are really fun - pistolbots keeping range but being a bit dim, railbots dodging, cannonbots being temptable(?) away from their target
* On tutorials again - it might be worth having a couple of sandboxish type tutorials, for weapons and especially buildings. As an example of why I felt the need for this was that the only time (so far) that I've built a control tower was in the last wave of area 3 when I had some spare money to experiment with. Before that I didn't want to risk diverting funds from turret building.
* Could the waves/levels thing be made a bit more obvious and maybe more prominently displayed? Plants vs Zombies does this really well. As a minimum perhaps you could just make the text bigger and display it as say "3 of 4" instead of counting down to 0. Or even just "Waves left"/"Levels left". In fact I initially found the word "level" here a bit confusing. How about "tides" (bigger than waves )
* Vehicles: I only tried the buggy so far, but I think it would be nice to press space (or whaetver) to get in/out and then have the vehicle controls the same as on foot (think of Halo or Halflife 2).
* Perhaps to cancel a build/sell mode you could have a "Right click to cancel" prompt on the bottom of the screen?

Overall though, as I've already said, it is excellent 
Did you suggest somewhere else that there might be a expanded part 2 version? I'm looking forward to it if so!!

Can I request some kind of (unlockable?) open-ended survival mode if one isn't already planned? 
(Oh, and maybe also a repair bot, somewhat like the money mite?)

@Inanimate: I guess I was trying to make it generally applicable to geography in exploration games... but maybe it's too specific to my project.  (Oh, desire paths - I kind of have them already, just using A*... it's near enough to the same effect )

@agj: I couldn't really give an expert answer as I haven't really read or studied much of the people who invented or practice it. This is just my interpretation... but I think you pretty much have it. The most interesting/resonant thing for me is the "derive"... maybe it's a bit like parcour without the jumps? I mean: You have a landscape which is defined by humanity - roads, boundaries, hubs, commuter routes etc - and so rather than follow the same old paths (perhaps dictated by authority, if you look at it that way) you do something else... something a bit more random and free.

I haven't read it, but Will Self and Ralph Steadman wrote a book called "Psychogeography" about walking from London to Dubai. They went via Heathrow (cheating) but still it seemed to be quite eye-opening.

Does that make any more sense?

I had hordes of zombie aliens spreading under my eyelids whilst going to sleep last night 

In a good way!

I know you aren't planning to work on it, but if you change your mind... (I wrote this before reading that post)

Regarding difficulty, I found level 3 to be the hardest (except maybe the last level, which i've only tried once). In fact, I was very close to ragequitting several times. I'll try and remember why:

* Accidental placement of buildings (wasting build points and handicapping myself)
Possible remedy: Allow incomplete buildings to be recycled for their full value? Make it easier to cancel the radial menu... not sure how. Does it ignore the selection if the mouse is outside the circle? Also could disable scrolling whilst the radial menu is up... I managed to do some weird stuff like place a HW zone outside the map.

* Conquering most of the map and then trying something risky and blowing it.
Possible remedy: It could be quite a bit of work, but perhaps some kind of save/checkpoint system?

Before tweaking the difficulty I would really suggest tightening up the UI and making it harder for the player (or just me) to screw up in a frustrating way.

I agree with Paul about the tweakable difficultly levels though. It would make it much more accessible and replayable as a bonus. Maybe just scale enemy hitpoints or something? I would certainly say the current difficulty should be the hardest, unless you want to add some kind of "insane" level above it.

For what it's worth, I would happily pay $5-10 for a version of this with a friendly UI and some more levels.   

This is really lovely. I enjoyed your previous albums, and this feels even deeper. Still loving the warm lo-bit sound

Oops, didn't notice you had a feedback thread. Like I said, this is great, I'm always really happy to see good indie RTS/tactics games appear.

standing still will only get me killed

I keep learning that lesson.... over and over... is that learning? It's really evil when the enemies make a dash for a control point that you thought was safe behind your lines