Game post-mortems

[VENGEANCE MX]

We've all had those projects where things went horribly wrong, and the occasional one were most of it went smoothly. Sometimes the games were received well by their audiences, and sometimes they were disasters.

What are your experiences with this? I'll write a bit about my own work, for starters.

Invasion
Brilliant new idea. Everyone loves it. Upgrade bug means that unit bonuses would be applied again every level, resulting in units killing the castle in one hit, moving at such speeds that they go straight out of the other side of the screen. Murder holes bug kill infinite numbers of infantry at once, resulting in 'all archers' being the only viable strategy. Castle aims arrows very badly. Made mistake of adding greyed-out units to the deployment menu which you couldn't actually unlock. Several saving bugs.

Invasion 2
Upgrade bug still there, but speed increase is removed. Arrow arcs become even wilder, greyed-out units still there, added useless ram unit. Archers still only viable strategy.

Invasion 3
Was received surprisingly well, due to amazing artwork from Kortex, and new units. Deployment keys refreshed browser for people on the Windows version of Internet Explorer (ie, almost everyone who played the game). Far too easy.

Invasion 4: The Last Crusade
Determined to regain the lost score from Invasion 3, changed the function keys to number keys, made much more balanced, artwork improved further, can't get into game due to broken preloader... whoops. Game becomes download-only, and is fixed 2 years later. Due to download-only-ness, the only version of Invasion that no movie portal's stolen yet.

Note: I still receive emails about the greyed-out units from people who're playing Invasions 1 & 2, on sites that stole them but cleverly overlooked the fact that they're not the latest versions.


Revolutionary RTS. Disaster. 6 months of work, 10,000 lines of code, and a full night's coding without sleep is wasted when Newgrounds audience don't read instructions.

[Squize]

Nice thread

On the subject of C3, aside from the slight bitterness , what went wrong ? What went right ? What with hindsight would you do differently.

Tell us about the highs and lows babe, the pitfalls to look out for when developing our own games etc.

Good, but I want to know much more

Squize.

[cadin]

I've written development notes for a few of my games over at The Pencil Farm:

Roach Attack

Back to the Garden

The Lake

I'm not sure my notes are very insightful, but I'm happy to answer any questions if anyone's interested...

[ihoss.com]

Commando 3 is an excelent game for being made in flash! I have always wanted to make an rpg, but I never managed (my old attempt ran way too slow). Excelent work, and too bad the NG kids are stupid.

[Squize]

Great notes there cadin, really good read.

I love this kind of stuff, the actual thought processes that go into making a game.

Squize.

[Squize]

Sorry about the double post, it's just that when you hit 4000 you get a gold avatar.

Found some old post mortem notes I was asked to write up for Heart Attack ( www.wedonicethings.com ), so I thought I'd share them here.
They're a bit dated now, so keep that in mind:

------------------------

Released: 14/2/2005

A game we produced in house at preloaded for the Valentine holiday.

I'd been working on a Dropzone clone for a couple of weeks in work as suggested to me by Rob ( Preloaded's creative director ) as he was a fan and I'd made no secret of it being one of my "Dream games" ( I'd always put off trying to code it, wanting to wait until I was good enough to do it justice ), when Rob suggested that we skin it with a Valentine feel and release it as a self promoting title.

I've got to be honest, at first I was less than happy about my hard as nails shooter being given a cute make-over. I did get very precious about it ( This lasted throughout the whole development, this was and is my baby ).

At this point there wasn't a great deal in the game, two layers of parallax ( Which I must admit to, still even now have the same bug in them that was there after two weeks ), the "Swarmer" baddie types ( Which became the little blue devils ) and the player movement / shooting. Now I was facing a deadline. The actual office deadline for it was for the Monday before Valentines day, the 9th of Feb.

Thankfully slotting all the different baddie types in wasn't too difficult, and in the end we had in all the ones we planned, averaged around three hours per baddie code wise, with roughly another hour on top of that per baddie to skin them up and make sure the animation was correct etc.

A lot of time was lost on silly things, I remember having a bitc'h of a time trying to get the pause mode working correctly, and it's still a bit kludgy even now. Same with the smart bomb ( For some reason some movie-clips just refused to either stop playing their animation frames, or would not start playing again after being restarted. Flash, you've got to love it )

We were constantly testing the game around the office on different machines as I wanted this to have more sprites than any other Flash game, and when we tried it on the 500mhz PC and Mac's with 3 layers of parallax and it still played well we knew we were on to a winner. For a while it was running at 60fps, which none of us really wanted 'cause we had visions of Flash8 actually being as fast as promised and the game just being unplayable, but even when dropped down to the more standard 31fps it still played really well.

Design wise, very close to the end of development the actual screen size was a cause for concern. Rob believed the play area to be too barren, esp. during the periods where you were just chasing after one baddie, and wanted the scrolling window reduced in size. I really dug my heels in about this, partly 'cause of the additional code, and partly 'cause I was concerned the game could lose some of it's "wow" factor if only running on a relatively small window. Rob had the excellent idea of dropping in the lampposts to break up the area, and it worked a treat ( For those of you who care about such things, there are 8 lampposts on a level, but those are reduced to only 4 later on as the amount of sprites hides the fact that I've removed some, and it freed up some all important cpu time ).

One other thing which I hope anyone playing it won't have realised is the actual timescale we had to produce the assets for the game. Both Rob and Chris ( A designer at preloaded ) were snowed under with other jobs, and I didn't recieve my first asset until around the 9th, which was our intended launch date !
Chris produced the title sequence images and worked on the parallax layers. As soon as I got the parallax images from him I was having to import them into the map editor and chop them up. From memory Rob didn't touch the sprites until around the Tues / Wed !It was a pretty high pressure time, as we were all aware that if we'd come out even a day after Valentines it would have suffered a hell of a lot.

As an indicator of how pushed for time we were, I don't think we had a name for it 'til around the 5th of Feb, and it was just known as "Bob" around the office. I really wanted "Cupid Stunts" as I couldn't think of anything more filthy ( Well, "Love Juice" was one of my ideas... ), but Andy ( Tech. director ) came up with "Heart Attack".

So that's about it on the actual development.We had the obligatory issues with the back-end stuff, for some reason as soon as I touch a php hi-score routine it breaks, and I had a 3 am session on the 13th ( I think Rob was averaging them for a week solid )

I wasn't happy that it launched with just the night tiles, and I managed to claw back a couple of hours here and there after it launched to tidy that up, fix a couple of bugs and add a cheat mode ( hold CTRL and press down your mouse wheel button ).

As to the game itself... I think it's the best thing I've ever done.
It marked the first time I had a lot of control over a preloaded project, and after all the fixes and updates it's pretty much everything I wanted it to be.
I'm so proud of the sheer number of sprites we've got running over a parallax background, it just flies along. I can't think of another Flash shooter which throws so much around the screen and keeps up a decent frame rate.

Aside from the technical aspect, I think it plays so well. I love old school blasters like that, and I must admit to being concerned that a lot of people ( The Nintendo generation ) would just switch off after finding it far too fast and difficult. I'm not sure if they did or not ( We had just short of half a million hits in the first two weeks ) but we got a lot of really kind words from people who actually remember dropping their 10p into Defender at the local arcade.

I'd love to re-skin it into my original vision for it, but Rob and Chris' pixels
are so great that it seems kinda pointless.

------------------------

Really don't like me semi-formal writing style that I used for some reason, but that's Heart Attack

Squize.

[Captain_404]

Crash Test
this was my first game, one of my friends also describes it as my most addictive too, personally, I'm not too fond of it because it was an indirect spoof of Trogdor

Super Penguin Bros
This was probably the best game I made, well, people tell me it's the most fun anyway. It's sort of like a bizarre cross between Mario and Megaman...I don't know. I eventually completely lost the file, otherwise this game would be MUCH better

BioNuclear
This is where I learned not to make things to complex, it probably would have been a great game if I had been able to raise the framerate over 4fps...this is the black sheep of all my games...

Dodge 'Em Up
The name of the game was a play of off "shoot 'em up", I think this is my favorite so far, it's probably the most polished too. The only glitch I've found so far is that on certain pages the music glitches up...

[cadin]

Great notes there cadin, really good read.

Thanks, I like yours too. (blush)

Seriously, this is great. I love this kind of stuff too.
Keep em coming.

[VENGEANCE MX]

Really want to read your notes, guys, and thanks for your interest in C3, Squizeh, will get round to looking through the thread and working on an analysis that does justice to Commando 3's awesomeness tomorr-later today. :P Stupid long homeworks. I was hoping to get some sleep tonight.

[alillm]

Yup great thread Will have to go through all the posts in more detail, but this kind of stuff is really good.

I’ve half written a full post mortem of the neon games, but as its not finished ill do a brief run down of each of my games (can't be bothered getting links).

Lumination
This was our first game as varstudios. The concept was great (I still want to make another one based on the same idea), but the major weakness of it was its overly simple, yet overly difficult gameplay. It also lacked in depth and didn’t progress much. If I do get round to doing another one it would be far more in depth with a good difficulty curve and improved gameplay.

Pixel Monsters
This was a weird one. I loved the concept, but hated the finished game. I still don't know how I even managed to finish it. After reluctantly releasing it on newgrounds expecting a terrible score, I was surprised to find that it did ok, and even generated a small fan base. Again I would really like to remake this, as I still think the concept was great. I probably never will though. Again, its main weakness was lack of depth and lack of gameplay. Its basis however was good, and it was the idea which caused it not to totally flop.

Viroidz
Our first shooter. Based off asteroids, but with some unique twists. This was the first game which I still liked after finishing it. The idea was pretty good, the gameplay was good, the difficulty was just right and there was enough to it to stop it getting boring quickly. The only thing I would have done differently now, is to make it faster and to add more enemy types.

Neon
This marked a change in the way I thought about making games. Instead of keeping to a boring formula and producing a average game, I just let myself build the game exactly how I wanted to play it while still sticking to a set goal in order to avoid going off the track. The outcome was a shooter which I enjoyed far more than viroidz and I felt like I had created something fun and creative. The only thing I wanted to do to it, was make a bigger better one.

Neon 2
...which is exactly what I did (well at least in my mind, many people disagree). I spent much more time on this one, fine tuning it and making it as good as it was going to get without me spending years on it. Got a bit carried away at one point and almost ruined the game through speed issues, but fortunately you guys helped resolve that and the game was a success. Some of the feedback I received however (mainly from newgrounds) was a bit discouraging, not because it was negative (most of it was very positive), but simply because people didn’t seem to get into it as they did with the first one. This left me unsure where to go next with my games and so I am now exploring some other shooter styles as well as some other ideas that i have had lurking around for a while.

Ali

[Son of Bryce]

Great idea Vengeance! I read a lot of Game Developer magazine so it's crossed my mind to write a post-mortem. In each issue they take a game and write about the 10 things that went right, and 10 that went wrong in the development. There's so many choices that you have to make during development and it helps to learn from others mistakes and successes.

I'm going to start writing some for my games.

[XareoX]

VENGEANCE MX alsome games i was playing you invasion games for minuites!!! (Not Hours Minuites) About 30 mins i like them they are addictive they are like a new kind of game invasion 1 is to easy and boring but ur last 1 is realy good Keep it up =)

[VENGEANCE MX]

Thanks Son of Bryce, but I'm afraid I'm gonna have to give topic credit to Squize, since it was his idea.

Alillm, you neglected to mention the fairly serious issue of Neon not working on Macs.

Nice thread

On the subject of C3, aside from the slight bitterness , what went wrong ? What went right ? What with hindsight would you do differently.

Tell us about the highs and lows babe, the pitfalls to look out for when developing our own games etc.

Good, but I want to know much more

Squize.

Was going to put this off another day because I'm absolutely knackered from homework, sleep deprivation, cross-country running and its resulting leg aches, but I noticed the topic needed a bump, so I decided not to delay it any more.

Things that went wrong (there are a lot of these, so, to maintain reader interest, I'm splitting them up into subsections:

Mac/Windows:

- Lack of proper right-click support in Flash. Most Windows users are accustomed to using a combination of Right-click and Left-click to play RTSes, so I suppose they're a bit alienated by the Ctrl-click method I had to introduce (although this is standard in most Mac games, since Apple only started making mice with more than one button a few months ago).
- Framerate issues (mainly on my computer). I'm used to Flash running horribly, since I'm on the Mac Flash Player, but most Windows expect framerates above 30. Pretty much all of the game's elements scale or rotate, so I wasn't able to turn anything into bitmaps, and the number of particles and movie clips flying around all of the time is staggering (seriously, I'm surprised it game out as smoothly as it did).

Project magnitude:

- Exporting time. The .fla file is somewhere in the region of 87.6MB, and the code for the main frame alone is 188 KB. Debugging took ages, due to the incredible export times (not helped by the fact that Flash's progress bars don't actually show the progress, they just flicker to random values and then freeze at the halfway mark).
- Filesize issues (or at least until I remembered you could compress pictures :P).
- Preloader problems. Due to the infamous issue of anything with a linkage identifier being loaded before the first frame is rendered (and since almost everything in the library had one), the preloader wouldn't show up until about 70% of the movie had been loaded (which takes a long time on slow connections).
- Damn that library. By the end of the project, I had over 700 symbols to trawl through, without a search field. Also note that Flash MX for Mac has no scroll wheel support, so I had to either click and drag the scroll bar (which was terrible for precision searching), or laboriously hold down the arrow buttons to scroll. I compromised by resizing the library to fill the entire height of the screen every time I wanted to find a symbol, and then slowly drag the scrollbar to the right letter of the alphabet, then use the arrow buttons to scroll up to the symbol if I couldn't get any closer with the bar.
- Coding panel. No scroll-wheel support on Macs, and it lags when I get more than 100 lines. No anti-aliasing, so the text is hard to read (although Windows users wouldn't mind this, you guys have it disabled by default).

Workspace:

- Screen resolution. I know that 1440x900 might seem positively luxurious to a lot of people here, and Flash isn't a full-screen pig on Macs, so I was able to arrange the workspace how I wanted to, but I always needed just that tiny bit extra screen space (found myself constantly resizing windows. Plus, the one bad thing I am prepared to admit about the OS X user interface - you can only resize a window from its bottom right corner. Made things frustrating, since I had to drag the whole window to the top of the screen before I could get at the resize button (if it wasn't all fitting onscreen).
- The dreaded Flash MX bugs. Flash MX doesn't save your workspaces properly in the Mac version (always gets windows in the wrong place, doesn't like it when you leave them half off-screen, doesn't save position or size of main window, just the panels around it). After a few minutes of usage, hotkeys stop working (this error still exists in Flash 8, never figured out why). When you change app and change back to Flash again, the grow box (thing you click and drag to change the size of the window) stops working. Scaling the window was a pain, since I needed to click off of Flash and back in again to fix the problem every time it occured.

Flash hating me:

- Code becoming unstable after a few thousand lines. Once I'd reached something like 8,000 lines of code, every time I added something to it there'd be a 50-50 chance of it breaking the game. Even removing comments would change things drastically. This was probably the toughest part of development, and the reason I'm not planning on making a sequel any time soon.
- Inexplicable errors. When the code became unstable, as mentioned above, units would flicker, the frame of the main timeline would rapidly change, bits of the interface would stop scrolling with the camera, you would no longer be able to control your units... it was horrible.
- The failed attempt to port to Flash 8 (I wanted to get its speed boosts, since the game was slow, and Flash 8 for Mac has OpenGL rendering, so the graphics are smoother). Just opening the file and exporting it (with the correct publish settings and AS version) resulted in units flickering, health bars disappearing, and just general instability and not-working-ness. It was faster, though, so I was a bit irritated by this.
- Flash throwing 'too many recursion' errors if I used the fill tool on too big an area with my flag designer (that's why the flags are so tiny).

Audio:

- First time adding volume controls. Just the interface itself was hard enough to implement properly (telling the bar to move to the right position, etc), but making individual sliders for music, sound FX, and voices work properly was hell.
- Getting sounds. All of the soundwork was done within the last 2 or 3 days of production, mainly involving hours of (for the most part) fruitless findsounds.com searches and editing them in Audacity.
- Getting the sounds to work. Flash doesn't let you use that many audio channels. Took me a while to realize that I had to remove all the whitespace in the sound effects so that the sound clip would be shorter than the firing animation loop.
- Finding voice actors. After weeks of dedicated searching, I found only one good voice actor (and I wanted one per unit). I couldn't think of original lines for every unit anyway, so he ended up as the global voice.

Misc:

- Collision detection. It was almost exclusively hitTest() and circle-to-circle, and I didn't feel comfortable using such basic methods on fast-moving projectiles.
- Saving. The disorganized way I stored some of the variables meant that the game was unplayable if you didn't have access to the SharedObject() (a school IT room issue). Made a small fix which only helped as far as stopping you getting to the main menu if the game couldn't save. Was meant to bring up the security settings dialogue, but debugging would require a 5-mile trek to my school between exports, so I gave up on trying to get this working.
- Sometimes units rotate the long way round (resulting in a snapping movement), due to my half-assed implementation of easing.
- Newgrounds idiots. Already mentioned them.
- People wanting your units to automatically attack enemies nearby them. Why not just make the game computer vs computer?

But what went right?

- Euphoria. Ever experienced when you find an incredibly simple fix to a serious problem? Just a simple if (i!=0) { fixed a horrible bug of where if you told units to line up, they'd do it about 20 pixels below the place where they were meant to, so if you told them to move somewhere, they'd be mispositioned.
- Formations. I love how the units form together so well and put themselves in line. You'd probably never get enough troops to discover this (since the map is 4 times bigger than in the original where I tested this), but if you tell units to line up in such a way that would force them to overflow off the edges of the screen (due to the gaps they leave between each other), the formation will compress to prevent this. I also loved when I had the pointless (and now-removed) circle formation, since it was beautiful (all the troops formed up with their rifles pointing in the angle of the circle they were at, so all sides were defended). I ditched it because it was useless in combat, since units could turn to face an enemy in about half a second anyway, the AI doesn't ever try to flank or surround you, and it actually made it so that less units were at the correct range to attack the enemy.
- Scrolling. The scrolling system is beautiful. At first, when I tried to add scrolling, I did it the 'proper' way, moving the map and all the player movie clips. But that let many collision errors, so I eventually just moved _root itself, and positioned the HUD in line with it at all times. I love the minimap, and how it locks to your mouse but won't go out of the sides of the screen (thanks to clever usage of Math.max() and Math.min(), and displays even Ion Beams and Troop Freezes in miniature size.
- Enemy AI. A huge improvement over Commando 2's 'move the unit somewhere random if it isn't doing anything, and if it's close enough to an enemy, attack' approach. It definitely feels as if you're playing against a human.
- Capture the Flag, and all the other assorted game modes. The AI I wrote for the computer always worked pretty much first try, which was a pleasant break from the usual "IT DOESN'T WORK, WHY THE HELL DOESN'T IT WORK?!" result.
- Writing the script in Apple's XCode. XCode is perfect. It has proper anti-aliasing, so I can read the code at size 6 point fonts (My Windows-bashing font diagram at http://www.birchlabs.co.uk/MacText.png uses XCode's text in the 'Mac' half of the picture). Plus, I can scroll! And there are anchors in the code that scroll me to any function I select from the functions drop-down menu!
- Releasing the game 10 minutes before the deadline in the most eastern time zone of the world, after an all-nighter which involved BEGINNING work on the campaign mode. I timed that well.
- The flag designer program. It's so good!
- The end result being the only proper RTS done in Flash (ignoring the fact that there's no resource collection or varying terrain types), and a great achievement for a 13 year old (although I'm now 14!).

So, in summary, more bad than good. Critically under-rated for how much work was put into it.

[axcho]

VENGEANCE MX, I just finished reading your postmortem, and wow, that game sounds like it took a lot of work. What's it called again? I never knew that anyone had made an RTS in Flash.

This isn't exactly a postmortem, but it might be helpful to some people. I made this game Braids, and it did really badly on Newgrounds, so I made a thread there to find out what was wrong with it. These were my conclusions, which I posted in that thread:

Braids

I never realized how difficult and confusing the controls were for most people. I can assure you that once you get used to them they are extremely powerful, but that doesn't really help you at all.

I will try to come up with a useful tutorial, one that explains the strategies of details of the controls. I guess the most important thing I can do is make sure the players know what they're supposed to be doing!

About the mouse control, well I know it's really weird, and causes a lot of problems. The ragdoll doesn't follow the mouse - moving the cursor 20 pixels to the right pushes the blue ragdoll's head to the right with a proportional amount of force.

Also, since the mouse button is used for crouching, I can't exactly put a menu button anywhere, or else it would be easy to click it accidentally. Instead the menu comes up when the cursor goes off the screen.

Eventually you learn to follow the ragdoll with the mouse cursor. It's pretty messed up, but once you're used to it, it's really powerful. Which, again, doesn't help you much.

Analog control sticks would be the best control scheme for this game. It would probably be a lot better on a console. But I think I'll experiment with controls where the ragdoll follows the mouse.

And let me clarify some things about the game. There is in fact an AI mode, because both players start out being controlled by the AI until someone drops in. If both red and blue (keyboard and mouse) have dropped in, then it becomes a multiplayer game. I'll probably make the AI start out better so it's easier to tell.

Also, you win by reducing your opponent's health to zero. It keeps track of the number of wins for each side, which you can reset by choosing 'new' from the menu.

And I'm still working on the sound. I'm working on programming some music that changes based on how you play.

So, once I get recorded games working, I'll release a movie of me playing against my brother so you can see what the game is supposed to be like! I think you will be pleasantly surprised.

After that, it's just a matter of making it easier for new players to get to that level of play.

I have to say that max0rz's comment was very enlightening for me. I realized that the game I was designing was structured differently than most games here on NG. It's designed for casual play, and it doesn't quite feel like a full game. I play it for the joy of feeling the *ahem* expertly designed physics, doing flips and handstands and such, setting up a big combo in slow motion.

Other games, even really bad ones, make you feel like you're accomplishing something. Like you're doing a lot, when you're only clicking as fast as you can, for example. I never bothered to even try to do that with my game. I didn't think of it. I just made a game that people who already understand the game and like playing it just for the fun of it, will like.

It was the first time I felt relieved and hopeful about the future of this game since its release, because it pointed towards a concrete direction for me to go with it. I need to show people how much there is to do in this game, because there is a lot more under the surface. I need to add music, to add recorded games so people can use it as an animation tool, as a social tool to show off and share their moves.

So, basically, I need to greatly improve the tutorial and controls, add the other features, and release gameplay videos and stuff to show what you can do with this game.

[fil_razorback]

- Damn that library. By the end of the project, I had over 700 symbols to trawl through, without a search field. Also note that Flash MX for Mac has no scroll wheel support, so I had to either click and drag the scroll bar (which was terrible for precision searching), or laboriously hold down the arrow buttons to scroll. I compromised by resizing the library to fill the entire height of the screen every time I wanted to find a symbol, and then slowly drag the scrollbar to the right letter of the alphabet, then use the arrow buttons to scroll up to the symbol if I couldn't get any closer with the bar.

Why didn't you make folders in the library ?

[Squize]

Excellent post big V

This to me is exactly what this thread should be, I imagine a lot of people reading this have just gone "Hell yes, I had that".

I'll try and go through stuff in there, but something that stood out straight away was the fla size ( 87.6MB ! ).
I've never had anything close to that ( Well, aside from the Evil9 video, but that was a 3 minute video ), and to be honest I don't know how you coped with that.

A little tip, a quick and easy way to do versioning control is to save the fla out every day with the date ( And store all the previous ones in a _dev folder ), eg sweetGame_2006_10_20.fla

Not only does this mean if you break something really bad you've only got to look at the previous day's working code, but also saving out a fla like that is the same as selected save and compress in the ide, it flushes out all the crap in the fla, reducing file size by a lot.

Sound also makes fla's bloat in size, so if you want the sound in there early for testing you can either import a rough mp3 version ( As opposed to the wav and getting Flash to mp3 it at publish time, which can be slow ), or just shove all the sounds in their own swf, and load that into the game at runtime, which is nice and simple and it means you only have to re-publish your sound swf when you add a new sample.

As fil_razorback said, libraries should make use of folders. I'm anal beyond belief with this, and always adhere to the same folder structure in any game ( I used to hate it when a designer imported stuff into a fla, 'cause there would just be files dumped in the library, I've really got ocd of the library ).

Squize.

[VENGEANCE MX]

Wonderful game, axcho, the only part that I didn't pick up within about 10 seconds was that you had to hit the enemy with your limbs (I was having lots of success hitting him with my head, but realized that it was my health going down when I did this :P). And since you ask, my RTS is called:





@Squize: The reason I didn't use folders is that I never assumed I'd be having a library this big. By the time I came to this realization, I had too many symbols to cope with and just decided to live with the mess instead of spending an hour organising everything. :P

I don't like saving separate files in Flash. My habit is hitting the 'Save' key combination every time I add a line to a drawing (one that I picked up from the days of Flash 5, which would crash with incredible frequency if I didn't do this). If something goes wrong, I generally just live with the fact that I've broken it, and try to rewrite that part of the code - it's easier that being organised. Flash MX doesn't have Save and Compress, but the 'Save As' trick you mentioned cut the filesize down to 30.2 MB. Incredible.

As I said, all the sound stuff was done in the last 3 days or so of production, so I didn't need to worry about getting it in early for testing. Loading external swfs isn't my cup of tea (would it even keep the Linkage properties?), but I'll bear it in mind in future. In my experience, loadMovie() hates people like me, so I'm afraid to touch it nowadays.

[UnknownGuy]

Interesting thread.

I just have a question about the ability of being able to go backward if you break something. Does everyone just make a new folder if you have lots and lots of external AS files?

How to deal with it I am not sure of(Copying the whole folder is cumbersome ).

[cadin]

I copy the whole folder.
But I don't make a backup every day, just when I'm making a major change to the game (like adding a new feature, or rewriting the way something works). I usually finish a project with 7-10 versions.

[axcho]

Wonderful game, axcho, the only part that I didn't pick up within about 10 seconds was that you had to hit the enemy with your limbs (I was having lots of success hitting him with my head, but realized that it was my health going down when I did this :P). And since you ask, my RTS is called:

Ah, Commando III? I will have to check it out sometime. And I'm glad you liked the game. I'm hoping to add music to it sometime; try listening to music while playing, it really helps.

I'll think I'll change the last instruction from "your hands and feet do damage" to "hit your opponent with your limbs".