next »

[Juan Raigada]

Ok, it's certainly been a while since I last posted here.

I promise for our next projects, since now we are a team and all, I'll try to keep a Devlog that doesn't stop updating halfway through.

But why am I talking about the next projects? Well, because we are coming closer and closer to the end of the development.  I am currently implementing the very last set of equipment to go in-game, and after that, we will be pretty much content complete (some gameplay polish left to do on the third level, but that will be all).

Which means we have started porting. After a week of tinkering (and learning about Mono trampolines) we have now a build of Heart&Slash running without any major bug (but some small ones) in XboxOne!

If 28 months ago (about the time I started the project) somebody had told me I would release a game on consoles, I would have thought them crazy as hell. We have moved from pretty much noobs to noobs with an (almost) finished and released game. It's a really strange feeling.

That the porting process has not been too painful makes me hopeful that in the next months we will be able to finally wrap the project and start looking ahead to new ones (yes we also need to sell the game, but we are working with some cool publishers here in Spain that will take care of that -hopefully-). Our current deadline puts us at being 100% finished with everything but release by December.

We will probably do some post-mortem once the game is locked, but for now suffice to say we have learnt a hell of a lot through this process. Our next games are going to be much simpler...  

[jctwood]

Very exciting, look forward to the post mortem

[Paul]

Crazy to see how far this has come, but it still has the charm of the original version! Will pick it up on PS4 instead of PC if possible as it's probably my main place to play at the minute

[The MegaMagic Dude]

But now, Juan, you must tell us more about that next project you just teased!

Congratz on all you've achieved so far with Heart&Slash. It was an honour and a pleasure to share the Kickstarter round table at PAD with you!

[Juan Raigada]

Very exciting, look forward to the post mortem

Yeah, me too. I mean, I don't know what I'll say, since I need to do a lot of introspection first.

Crazy to see how far this has come, but it still has the charm of the original version!

I hope this is true, it's hard for me to say anymore.

But now, Juan, you must tell us more about that next project you just teased!

Not yet, not yet. We need to wrap this first. Plus, I don't want to say anything until we have prototyped some and are sure we will develop an idea. Right now all the preliminary design is just being done on paper.

[Elideb]

One more update! Let's try to keep on track.

This week we were featured in SONY's China presentation, as part of their push for indie games in the country . We are really proud of that, even if Heart&Slash is seen for only a few seconds and the trailer needs an update desperately . We were displayed alongside our friends from Anima: Gate of Memories and Ziggurat.

https://www.youtube.com/embed/x1GXcWBbkHM?rel=0

We are now four equipment pieces short of reaching our quota for level 3 and Juan is working on them. Next week we'll do a review/bugfixing pass on all of them, so they should reach the builds soon. We have some funny parts in there, with weird abilities and trade offs. This is one of the greatest goals we have left, so once it's over, it'll be time to review the Space Lift level and crazy amounts of bugfixing.

On less content related issues, the porting we did for Xbox One forced us to move the game to Unity 5. There are still some issues related to that: shaders compiling with errors and misbehaving physics, mostly. Once those issues are solved, we'll merge this development branch into the main game. It won't be easy, since a few meshes will have to be replaced, but it seems feasible, albeit daunting. But we have some other priorities before that.

One of those priorities is getting the game ready to work in a (for the time being, unnamed) Chinese Android console. We are close to getting it to work, but there might be performance issues. If that's the case, we'll need to optimize assets and levels. We seem to be doing alright regarding RAM and CPU usage, but we are so far unable to get much out of its GPU. Any optimization we do in there, of course, will be moved to the main development branch, if possible.

As things move on, we'll let you all know what's up.

[migrafael]

i remember seeing this a while back. COngrats on getting on the new gen consoles! 

[Elideb]

Aaaand once again, we let too much time pass between updates.

Quick recap of the project status:

• All equipment, pending some debugging and improvements, is in the game. YEAH!
• 3+1 playable characters are in: Heart, the Prototype (very hard mode), Slash (vanilla mode) and the secret costume
• The game is more or less working on 3 of the 4 target platforms (we haven't got a PS4 devkit yet), albeit memory requirements are still too steep.
• The first, temporary version of all dialogues is in the game (English only)
• The credits screen is in, in a placeholder-y state (if that does not say the end is nigh...)

And the work, more or less, still missing:

• A couple of music tracks
• A good deal of sounds
• 2 extra characters from our Kickstarter backers, one of them half-way through
• A little more tuning of Copgore and QUASSY's fights
• A lot of tuning on the Spacelift level
• Text revision and translations
• And lots of tweaks, checks, bugs, tweaks, checks...

So, what have we been up to lately?

The equipment has been recently finished and we expect to have the extra characters finished in 2-3 weeks.

We also introduced the unlockable Slash character. It plays like normal, but it does not get any of the unlockable stats upgrades Heart does (extra hearts, resistances and strength). It's an option to replay the game with the same difficulty it had the first time around, with all the unlocked equipment.

A true brother of metal

Extra characters mean that we need a way to control which cutscenes play for which characters, and which content they have access to. That's required a little rewritting of small parts of the game and some re-configuration -and a bug fixed, as a bonus-. Now that's out of the way, only the new characters remain.

Regarding the ports, we are doing some tricks in them, which we'll move into the PC release, as options. The most interesting is dynamic resolution (adaptive resolution? Is there a common name for this?). It seems to be the norm right now, but the first time I read about it (WipeOut HD) I thought it was a great idea. WipeOut HD's version was more elaborate, of course, but ours seems to do the job. Basically, we check the framerate and when we detect it is too low (<25fps), we reduce the resolution we render the game at. If it recovers over a higher threshold (>=30fps), we increase it. Since we keep the UI resolution fixed, we've run into some issues when converting world positions to UI coordinates, but we're working on that.

We're tweaking/revamping the fight against Copgore, too. That's the boss we're less happy with, so it's being changed considerably. New attacks, movement patterns, probably changes to the camera, and a fix to the clipping in its entry animation.

It loves the smell of burnt metal in the morning

And that, plus fighting with each console's special characteristics, is our life right now.


As always, we'll try to update more often, comment on the stuff we're doing in more detail, be more active, etc.

[Moth]

That dynamic resolution idea is brilliant, good job implementing it.

This game is definitely one of the best looking indie games ever. I'm not sure I've ever seen a more captivatingly beautiful 3d indie effort. The style is wonderful and the game looks really good, I can't wait to play it!

[ACOM]

I have realized that you are here! Good place to be

When does the game goes out?

Cheers!

[Juan Raigada]

That's going to depend on the publisher that's going to help us with the marketing. Probably we'll release simultaneously on PS4, XBoxOne and PC somewhere in February/March.

But I expect a 100% content complete and mostly bug free version in Steam somewhere in October, though. By December we want to be working 95% on the next project. at this point 95% of the work left to do (aside ports) is bugfixing... the only problem with this version should be the price (it will keep the early access -higher in our case- price until the multiplatform release).

[Elideb]

Let's explain a little bit¹ about the dynamic framebuffer thing we've implemented for Heart&Slash, in case I need to refer someone in the future.


Original idea

As I said before, we got the idea from WipeOut HD. Sony's Studio Liverpool (previously know as Psygnosis) used this in WipeOut HD, Fury (PS3) and 2048 (PSVita). As soon as they detected a frame was dropped, they would reduce the horizontal resolution of the game. Then, the PS3 would use its hardware horizontal upscaler to smooth the output image. That kept the framerate always close or over 60FPS, without the player noticing much had changed, as it usually happend when the ships were moving very fast, or lots of ships/effects were on screen. This required disabling v-sync, however, so the ocassional dropped frame would be less painful.

If you want some more info regarding these, Digital Foundry had some articles about it:
http://www.eurogamer.net/articles/wipeout-hds-1080p-sleight-of-hand
http://www.eurogamer.net/articles/digitalfoundry-wipeout-2048-tech-interview

Of course, Studio Liverpool were not the first ones to do it, and neither were the last. Nowadays it's very common for all sort of games to use a render target of non-native resolution for the game. This render is later upscaled to native resolution and merged with the UI to generate the final image the player sees. Still, not many of them change the resolution of this render target during the game.


Our particular case

We have found ourselves with the task to port Heart&Slash to an Android console. It is quite a powerful machine, using a quad core ARM chip and nVidia's Tegra X1 graphics processor. Add 4GB of RAM, and it turns out to be comparable to the desktop computer I had 3-5 years ago. But, although the machine is no doubt able to run our game pretty well, when rendering at 1920x1080 the framerate can drop below 20 frames per second in the biggest (not very optimized) rooms. So we'd just run at a resolution of 1280x720 in this particular hardware, disable some of the most demanding effects and be done with it.

Unfortunately, the system locks whatever resolution it is configured to use, so in a 1080p television, the game MUST run at 1080. When we discovered this, I remembered the WipeOut articles and re-read them. Unity offers the option to easily render a camera into a texture (we already did this for the minimap), so it was feasible. We only needed to verify we could control the resolution of the render texture, and turns out you can.

Our first, crude implementation simply reduced the whole resolution of the texture, in 8 steps, down to half of the original resolution we were using. This produced very dirty images, with very noticeable pixels. Re-reading the articles, I noticed Studio Liverpool only touched the horizontal resolution, which makes a lot of sense. Using 16/9 aspect ratios, any reduction in vertical resolution will be a lot more noticeable, since it means lots of vertical detail, which is already scarce, will be lost.

The next implementation allowed us to adjust how low we let the resolution go in vertical and horizontal, independently. Configured to .5 in horizontal and 1 in vertical, we now can go from 1280x720 down to 640x720, in 5 smooth steps.

1280x720 VS 800x720
Click to see the original images

That's good enough for us, and we might leave it at something like .6 in horizontal, 1 in vertical, to improve the quality of the final image.


How to do the change

So that's for the options we have for resolutions. Next, we needed to decide how often we allowed changing the resolution, and which circumstances would trigger a reduction or increment in resolution.

Originally we used our basic FPS measurement widget², which checked the framerate over 10 seconds and dropped the seconds with the most and least frames to calculate the average³. Using averages meant a very bad second (loading a new room) would drop the resolution by a lot, and it would take long to recover from it.

So I reimplemented it to use the average of the last 100 frames, shaving off the worst and best frames (5% of them). It is faster to calculate and manipulate this way (a fixed size array in which I replace a value every frame, plus a little logic to check if the frame length was/is taken into account for the average). With this change, we check more often and accurately, so the resolution change reacts faster to worsening and improving conditions.

Things to take into account (mostly related to Unity)

• When changing global resolution, make sure you set your camera´s aspect ratio the the correct value. Otherwise it will take the texture's aspect ratio.

When aspect ratios go wrong

• In the editor, don't use the screen resolution, but Screen.width and Screen.height. These values contain the size of the game render area Unity is using.
• The rendered texture must be displayed in a final camera, filling it completely. We use a NGUI texture which is painted below the UI, anchored to the screen borders.
• Define the lower and upper framerate limits with a little bit of room: 30/33fps is a good value. Otherwise, increasing the resolution would get you back below the lower threshold, and you'd keep switching resolutions endlessly.
• Adjust graphical effects, maximum/minimum resolution and FPS thresholds per platform. A PS4/XboxOne/PC can keep at 60fps easily, with lots of effects; the Android console cannot.
• Some platforms impose limits in the minimum framerate the game must run at. Set your thresholds way above this and configure the resolutions with care.
• Give PC users the option to enable this, but never enable it by default. PC players tend to be more picky about stuff like this.
• No matter how low your game render's resolution goes, you can keep rendering the UI (your effects, too, if rendered by a different camera) at the highest available resolution.
• All your code used to convert world positions into camera coordinates (Camera.WorldToViewPoint) must change. The resolution of the camera does not match that of the UI, so a pixel X, y in your game camera has little to do with a pixel x,y in your UI camera. Either use ratios instead of pixels when converting from game camera position to UI camera position, or use another camera to request the transformation.
• You need to release the render texture and then recreate it, before changing its dimensions.

Code:

this.renderTexture.Release();
this.renderTexture.width = this.currentResolution.Value.Width;
this.renderTexture.height = this.currentResolution.Value.Height;
this.renderTexture.Create();

The end result

Thanks to this forced resolution, the game now runs smoothly on the Android console. It rarely goes below 40fps at 720p, so we've never seen the dynamic resolution kick in. The real treat we've got is that now we actually have control over the resolution the game uses, instead of letting the console decide that for us. And, since we have gained some margin, we can enable more effects in the console, like bloom, low quality shadows or ambient occlusion.

Comparison from 843x949 (far left) to 1687x949 (far right). The differences in the cel-shading are quite evident.
Click to see the original image

Closer comparison, 843x949 (left) vs 1687x949 (right). Notice the loss of detail in Heart's face and the C.O.P.S. statue.
Click to see the original image

Things to improve

• The cel-shader we use is heavily affected by the resolution of the render target, so as resolution lowers, lines get wider. This is the most visible effect and that which makes the final image look worse. We'll need to either try different algorithms, or apply the cel-shading to an upscaled version of the texture.
• Using antialiasing on the final image could help mitigate the dirty look the lower resolutions have, compared to the higher ones. But it might affect render times more than simply downscaling less...
• Probably lots of stuff, adjustments, fields of view...

UPDATE: A short warning. We've noticed that using this technique, at least with our current settings, produces banding in the render texture. It is specially glaring when composed with the fine detail Ambient Occlusion adds.
Some render texture configurations fix this, but adding problems of their own, so we're still investigating. For the time being, we'll probably only use the render target when Dynamic Resolution is enabled, instead of having it always active.
We're still using Unity 4.5, so this problem might have been solved in later versions. But we cannot change until the port to the Android console is complete...


¹ Well... it got out of hand and it is now quite the post.

² Don't rely on Unity's FPS stats.

³ We still displayed the average for the last second to the user.

[archgame]

I backed this game awhile ago and had no idea it was on tigsource! Your latest post was really great and I'll be taking the next couple of days to go through the entirety of the log, if the other posts are even half as informative as your last one this place is a gold mine.

I honestly sat here for a couple of minutes trying to notice the difference between the screenshot comparisons and you are right it really only shows up on the statue, which was so far away anyway that it didn't really matter, besides the games always had this grainy voxel vibe so the deresoultion actually fits in with the overall aesthetic.

[Elideb]

I backed this game awhile ago and had no idea it was on tigsource! Your latest post was really great and I'll be taking the next couple of days to go through the entirety of the log, if the other posts are even half as informative as your last one this place is a gold mine.

I honestly sat here for a couple of minutes trying to notice the difference between the screenshot comparisons and you are right it really only shows up on the statue, which was so far away anyway that it didn't really matter, besides the games always had this grainy voxel vibe so the deresoultion actually fits in with the overall aesthetic.

Thanks!

I'm not sure we have many in-depth articles around. I plan on changing that, but they take up a lot of time. And we are short on that, unfortunately.

Regarding the screenshots, the reduction in quality is noticeable when playing. Even if you only notice it when things get really bad or when the camera is standing still, it is something we need to improve. As I said, most of it comes from the cel-shading, which draws x pixels, instead of distances/ratios of the image. We'll probably need to calculate how much to paint differently, so the width is consistent between resolutions.

Juan has always believed that the cel-shading looked better at 720 than at 1080, because of the wider lines. This is a good change to see if we can match the look in the highest resolutions. Although this is a complex problem.

Worst case scenario, we upscale the texture to 1280x720, apply the cel-shading there and then adapt the resulting texture to the proper camera resolution. This kind of things get really messy, so let's see what we can manage.

[Elideb]

We've been busy this past week. Juan has fixed lots of bugs, drawn lots of icons and integrated the final playable character in the game!

Out with the old, in with the new!

This new character is very different from the others, so apparently it's been hell to integrate. And, for the time being, it's gameplay will remain a little simpler than the rest. But it is fast and plays incredibly well in the air. Let's see what we can get out of it.

The reason why I can't really judge how hard it's been is because I've been banging my head against the dreaded final stages of multiplatform: achievements and other platform specific services. But it's paid dividends. Now we have a decent and relatively painless way to integrate achievements and other stuff. The achievements have been integrated into the Android console with surprising ease. The most tedious part has been adapting the icons to the required dimensions and adding them all through the management web.

Who's the boss of you? I am the boss of you!

I've used the opportunity to properly sort achievements/quests (tip: always include int and string ids for your achievements; each platform uses what it pleases). That was one of those small, low priority tasks that always get pushed to the future "when we've got the time". It really bothered me, because the quests screen seemed much more chaotic than it should. It felt very unprofessional.

There's lots of small stuff like this that, in the end, never gets fixed because of a lack of time. And these little things add up, so the less we leave for the end, the better.

We'll see how this all works when we integrate Steam and GOG's libraries!

Edit: Removed the last shot, which could be a little spoiler-y 

[Elideb]

I updated the post about Dynamic Resolution with a warning about side effects we've detected:

UPDATE: A short warning. We've noticed that using this technique, at least with our current settings, produces banding in the render texture. It is specially glaring when composed with the fine detail Ambient Occlusion adds.
Some render texture configurations fix this, but adding problems of their own, so we're still investigating. For the time being, we'll probably only use the render target when Dynamic Resolution is enabled, instead of having it always active.
We're still using Unity 4.5, so this problem might have been solved in later versions. But we cannot change until the port to the Android console is complete...
sup] We still displayed the average for the last second to the user.

By the way, is it me or have 343 Technologies copied our idea? Of using this widespread technique as a marketing point, not the actual idea, mind you . That article must be the web page with the most appeareances of "60fps" and "up to 1080p" in the internet.

[MaartnBoot]

This is looking really interesting :D Digging that style!

[Elideb]

Hey! Even if we're not very active in here, the game keeps progressing!

Pretty much all the audio is already in the game, all tracks are too and we are in full getting things closed mode. The full port to Unity5 has been less painful than we expected, with the physics behaving pretty well. The Xbox One version of the game is already running, with achievements. There are lots of things left to get the platform release ready, but we'll get them done in the coming weeks.

There's still some more optimization to do, too, and bugs to fix, with a lot of our attention also dedicated to that.

And then, there's this:

And more languages to come!

We do that funny trick of displaying the IDs of the strings so that localization testers can quickly find and fix incorrect lines. We had to standardize all the scripts outputting text, so this could be done without blood, sweat or tears. But once it's been done, it looks as if it required no effort.

Editor side, we have setup our system to export and import translation packages, supporting full (new or heavily modified English text) and review (<10% and <10 characters changed) batches. Thanks to my past experience working on this stuff for EA, things have gone much more smoothly that I could have hoped for.

And, thanks to all this, the game looks pretty good in Chinese!

Spoiled brat.

When I have the time, I'll try to talk a little more in depth about the translation pipeline. I think that could help others, when localization time comes.

P.S.: Now that I see that, we might need to make Chinese dialogue write slower. We're currently using the delay between latin characters to separate Chinese kanji, which is probably not the right thing to do...

[Juan Raigada]

Ok...

Maybe you noticed something about the thread icon .  Heart&Slash is done now... Doesn't mean we will stop working on it (there's still porting and bug fixing/polishing until we launch), but it does mean all creative/design/hard decisions are taken and that the game is basically in its final form. 

It's been quite a trip.

This is just a pre-holiday post that I wanted to get off my chest (that thread icon HAD to change), but starting next year I plan to post a series of small post-morten explaining some development/business decisions we made, and what went wrong and right with them. Not exactly game post mortems, since we want to wait to release to see how the game is received before doing that, but stuff more process-based. Man, did we learn from this!!!

Also, I'm very aware I dropped the ball on this devlog a while ago, once the project grew. The energies I was pouring here were diverted to team and project management (and honestly, production is a much boring phase that early prototyping/implementation. It's early in the projects when the cool stuff happens). Can't really do much about that except trying harder to keep it updated on our next project.

Anyway, happy holidays to you all. I'll come back in a week or two to start being philosophical about things...

[Juan Raigada]

HEART & SLASH - PRE RELEASE POST MORTEM 1

PERSONAL NOTES... TWO YEARS LATER

Hey all...

Sooo, so much for a week or two. This is the first of the promised postmortems.  This one is focused on my personal thoughts and insights after having gone full time/ low money indie for two years. It is thus very personal and focused on my experience, and it does not necessarily represent the thoughts of others working on the game.

I'm writing this as I'm doing a H&S build to fix some of the bugs we found in the pre submission to one of the consoles we are porting to. The game is still not finished (preparation to release takes time, but the game itself pretty much is, and has been for a while).

Two years ago, almost to the exact date) we launched the H&S kickstarter, the success of which implied we would go into full time development. It was only three months earlier that I personally decided to take the project and finish it as a complete game instead of letting it be just a prototype. So it was about 27 months ago when I took the decision to switch careers from film into gaming.

That's a looong time. It is even longer when you are over thirty, have no real previous experience in the industry you are switching into, and your savings are not necessarily big (almost inexistent).

I don't know how others approach starting full time development, but in my case I was always clear this would be a marathon. I had very little expectations of immediate short term success, and the plan was always to start building up experience little by little. Never to rely on a big hit to keep marching on. The idea was to never put too much at risk (except our own income) and push onwards as far as we could with the game. And then release when we felt we either couldn't go forward anymore with our current experience, or because development became unsustainable due to financial constraints. Fortunately for us, several developments allowed us to not stop until we felt we had accomplished far beyond the original goals. Deals with foreign partners and local small publishers meant we had extra sources of revenue (nothing big, but it helped) and support to release the game in consoles without too much pain (although there's still quite the pain involved, tbh ). But this also meant the game would take longer to be made, and that in turn implied I had to return to part-time teaching to minimize financial risk, which itself pushed the release back a little bit more.

At the end of the process, we have shifted from a 14 months estimate (after the kickstarter finished) to a full 27 months of development, almost doubling the original estimate.  Not too bad, really, specially since we were newbies and at least 3 of those months have been taken up by porting that was not originally in the plan, probably more, although it's hard to measure precisely).

In part, that has been part of why I made much, much less frequent updates here the last year. Not only did I start to need to use the time I used to communicate to rest and disconnect from the project, it also was because much of the process was no longer creative, just pure, not that interesting, development. Many of the fundamental insights we could share we had already shared, and I didn't want to spam with too many meaningless updates (we had Steam for those interested in following more closely).

I'm fully aware PR is part of the work of an indie developer, but I didn't think PR during such a long period was necessary. I honestly was scared people would lose interest in the game if we just kept showing things with no real end date still in sight. Also, we now are working with a very friendly small, local indie publisher, and they are going to take care of that part of the project and I don't want to step too much on their toes. This blog and community is very dear to me (one of the reasons I jumped into this), so I will keep clean of marketing (and thus why I'm writing all this). The plan is to step up on the PR front from now till release, though

Another reason, I could not maintain the energy levels required to keep the update ball rolling for so long a period. At the beginning of the process, excitement and novelty can do much to keep you in a state you can work 12+ hours pretty much every day and not suffer it. But this is not sustainable (nor desirable, honestly) for such long periods. It becomes even harder when you also want to have a life external to the industry (I got married 16 months ago) and age doesn't help either (I'm in my mid thirties, having come somewhat late to the party).  For the last year it's been only two of us working on the game, and I wanted to keep my energy focused on developing. Everything external to it took a backseat. And that really helped us push the game a long way forward.

If the above sounds sort of down beat, or not positive enough, to you, make no mistake. It's definitely not. I'm actually extremely happy of where the project is at, how we are working as a team, and excited about our next projects (can't wait to start prototyping). Yes, we got delayed. Yes, we were non-optimal in many decisions... But we were also learning to design, create, iterate... We had to navigate development processes that I had very little to none experience with. The fact that we had emerged at the other side with a game that (despite the usual misgivings as a creator towards the creation we all have) I am really proud of, and with clear plans and intentions to do more projects after that. And, most importantly) financial stability to do so at least for one or two projects, no matter what happens on release of H&S.

It's just that game development is now my full-time job. And while you can (and have to, in the case of games) be passionate about your job, there's a big difference between developing as a means of sustaining yourself and developing as a hobby. There's magic here and there in the process, but that magic is far from day to day. Specially towards the end of the process, when you are no longer amused and surprised by how the elements start fitting together on the game, and are instead fixing bugs and polishing behaviours of elements you have seen countless of times. The day to day is a struggle full of (very, very satisfying) personal growth and learning. But it is also tiring sometimes, and, most importantly, it stops being something special and becomes what you usually do. The approach and focus had to change (in my personal case, I think people cope with this differently) by changing from a full steam ahead approach to a more measured, maybe more thought out, approach. It was about not burning out on the game so we could keep it fresh to the players even though it was obviously far from fresh to us.

I think I prefer it this way.

See: Three years ago I had NEVER worked on games or even remotely close to the industry. I was a 31 old professional wanting to change into a different industry (although, to be fair, I had dabbled here and there with Unity and I had studied CS while in college 8 years before that). It was exciting and frenetic at the time, all the things happening as the project got started. But I also had nothing. I could not really define myself, I could not answer to the question What do you do for a living?. For all the possibilities, I also felt somewhat lost. But now? Now I have finished a game along with some great partners. I have worked and learnt (a little of) how the business and the industry works. I've met other local developers, and I have been accepted as one of them. I can now say that I make games, and that feels really, really good. It's just not something excited about any more. I feel happy, and happy feels much better than excited, or at least it does so now.

...

So, that's it, my personal postmortem (or a postmortem of what happened to me during the project). Sorry if it feels a little bit like  a mind dump. It somewhat is (although I edited it) but this place is were all this started and I wanted to share. Sometimes projects and development is not just about the product...

And again, all the above is exclusively my personal take on all this. This is not my project, but the project of many people and we all own it creatively (and in other ways). I'm pretty sure my experience is not exactly similar to that of others, specially since we all started in different places.

We will do some more technical/business oriented mini postmortems before release (we'll see how many).