Saturday, May 9, 2009

A Long History of Katawa Shoujo

This is something I've been meaning to do for a long time, and every time I look at the sidebar of the blog I'm reminded of it. The dev blog was originally intended as some kind of semi-official chronicle of making KS, to be saved for posterity, and I became the semi-official chronicler. However the blog covers only the period from its inception, many of the posts are not really that interesting in retrospect and the short history on the website is rather curt. So here, a long history of Project Katawa Shoujo:

The seeds of KS were sown almost ten years ago, in the Comiket 61 held in December 2000. Doujin circle Zettai Shoujo and its artist RAITA released a Nausicaa of the Valley of the Wind doujin called Schuppen Harnische, centered on the amputee warrior princess Kushana. On an omake page RAITA drew concept art for a disabled girl dating sim which he dubbed Katawa Shoujo, drawing the five main heroines and writing short descriptions in the margins. Notable is that there were two other omake pages next to it, describing two other ideas RAITA had. Why these two became forgotten and KS survived, time and coincidence will tell us.

Fast forward six years. A western fan translates and colours the KS omake page (but not the two others), becoming not only the person who sets the project moving like the first stone of an avalanche, but also unknowingly becoming one of the biggest influencers to the visual appearance of the game, singlehandedly deciding the colour scheme of the school uniform and the hair and eye colours of the main heroines. The exact date of this event is unknown, but the doujin page is posted on 4chan /a/ on the 4th of January 2007. In retrospect, what happens next is the most incomprehensible link in the improbable chain of events leading to what project KS is now. For no apparent reason at all, the entire /a/ board goes batshit over the concept page and the thread starts filling with ideas about what the game would be like and calls for actually making this game a reality. Several threads about KS are made and filled, until a moderator stickies one of them, a rare occurrence on 4chan. At this point the momentum of Anon is too big to hold back, so they collectively decide to go forward with this idea of making a game based on nothing but the single omake page RAITA did. A person who we will call P arranges a development forum, since 4chan is a rather poor development platform due to the impermanence of any content posted there. For the following weeks, this forum works like a Petri dish of creativity, as swathes of completely unorganized raw ideas are born, fight each other, evolve and die. The most active users, P included, attempt to structure and organize, with varying results. Most of the heroines gain some semblance of personality and names, and the setting is explored a little bit though it eventually sets into a very very basic Japanese high school, except with disabled students.

The interest of a large portion of anonymous eventually withers and dies, as it always does, and the most active users now try to figure out how to proceed. Core Group, a de facto leader cabal, has been formed, aiming to give the project some structure so it actually could get some work done. They try to trawl the visitor base for contributors of sufficient quality, but nothing seems to be moving forward much despite good effort. There are arguments between people who want KS to be kind of an open, free-love community project where anyone could contribute, and those who want it to be more in the style of a closed developer group who make the game from start to finish. Eventually infighting and clashing egos lead to drama and P leaving the project, with the leadership left in the hands of cpl_crud and someone who we call T.

At this stage (April 2007) almost the entire contributor/userbase of KS has been demolished because of the lack of progress, disillusionment from the regulars and waning interest from the fickle anon. The entire project is in severe danger of dying off. Crud decides to start anew, and arranges new development forums on another webhost. These forums literally start from scratch and the few remaining devs move there with only the ideas that they have collected and distilled over the first four months of project KS. They deal the responsibilities between themselves, ending up with crud writing Hanako's story, A22 writing Shizune's story, someone we call S writing Lilly's story and Hivemind becoming the head editor. K will work the character art with F doing background art and being the art director. T becomes "director", essentially starting to learn to use Ren'Py, the VN engine of our choice. I rejoin the project in June after coming back from hiatus, becoming the Rin path writer and eventually Hivemind agrees to write Emi path in lieu of the editor mantle. Thus, the first iteration of the Four Leaf Studios has been assembled, and we take on that name for ourselves, as a kind of a joke and a reference to the origins of the project. Even though the name is there, at this point we hardly could be called a studio, as nobody has any experience in VN making and the professionalism of the endeavour is laughable at best.

The summer 07 passes in learning the ropes and trying to plan the game. Even at this early stage, the project plans to make a full length visual novel, which is crazy ambitious to say the least. To this day I am not sure how the hell anyone could think it was a good idea to make a amateur full length VN on the first attempt at the medium. But there we are, going along with it anyway. Unfortunately A LOT of bad decisions are made during this period due to inexperience, and this will be costing us dearly in the future as you will notice. Ignoring the importance of good planning (I (unfairly) blame crud for this, as it's one of his strongest characteristics as a creator, even to this day), the dev team starts forging half-blindly ahead, churning out text and pictures which are then assembled in Ren'Py into the first playable version of KS. Many of the side characters are created in this period, and the main cast gains more personality. Many, many things from this period form the concrete base of KS as it is today, and the dev team refers to them as "the stuff we want to change but we can't". The playable alpha demo of the first day of KS, called Grid1, is leaked to 4chan in August and it generates a slight renewal of interest for some time. NicolArmarfi joins the team, giving us a composer who can produce an original soundtrack for the game.

Sometime in the early Autumn 07, the first occurrence of what will later become some sort of a sad pattern, comes to pass. The dev team realizes that pretty much everything so far was a horrible mistake caused by lack of foresight and planning, and is going to lead to problems of apocalyptic scale unless everything is scrapped and the writing work started again. Unwilling to waste the work, the dev team decides to salvage what it can, and I proceed to stitch the existing writing into a newly structured framework of my design, with the writers filling in the holes. The basic structure of KS is formed here (and it's going to persist all the way to the final, full release). The first week, Act 1, is shared between all paths and written in collaboration between all writers until it splits on the last day into the path stories of the girls, written by their respective writers alone. After the work on Act 1 is finished, the writers move on to their path stories. In late Autumn I realize what a shitty and hasty job I did with the restructuring, and do it all over, this time all by myself. Thus we are up to the third iteration of Act 1.

Delta joins the dev team in September and starts to learn Ren'Py and Python from scratch, planning to code a proper UI for the game. Meanwhile, K quits the project, unhappy with his own capabilities and S gets into a serious accident, leaving Lilly without a writer and the project without a character artist. However, December sees Suriko joining the devs, replacing S as Lilly's writer. The art proves to be a bigger problem. Though we don't know it yet, throughout the autumn, winter and spring we are going to chew through a number of artists, all of who join enthusiastic but end up quitting the project for one reason or another. This period is indeed marked by the awful problems the project had with artists, as not a single background image had yet been finished (in eight months of working!) and the lack of character artist spelled uncertain future.

Despite this, the project moves on in other fronts. Through the spring and summer 08, the writers finish up the drafts of the path stories and delta increases his proficiency with Ren'Py, customizing the engine itself, working under the hood and tweaking the UI until it starts looking something like the highly tuned version you see in Act 1 release. A group of editors, including Silentcook, Losstarot and Kagami, join the team because there finally is something to edit properly. We have them and a small group of proofreaders go over the script of Act 1, and consider releasing it with the old art by K (but end up not doing so).

During late spring, a ragtag group of artists who are friends with each other offer themselves on the altar of project KS, and thus moekki, Ambi, weee, gebyy-terar and kamifish become the people whose handiwork you see in the character and CG art of KS. They split the workload in a similar way we writers do, giving each character to the sole responsibility of one artist. They spend a few months practicing getting their art to be sufficiently consistent with each other and decide an art style for the game (moe-moe in the fashion of Leaf's games). In summer, the real art production begins, starting from the sprites. At some point Nicol quits project out of boredom, as he considers there is enough music. T and F are fired from 4LS. Realizing that the lack of backgrounds is going to be a problem, we call out for contributors who could get appropriate photos we could base the bgs on. Luckily Yujovi and his trusty camera come to help, and soon we have more photos than we need for two games. Me and crud visit Japan, conveniently enabling us taking photos that require native flavour such as the festival backgrounds to complete yujovi's work.

Since the entire script of the game has now been finished, it's time to look back and see what we got. AUGH. We decree that the script of Act 1 is of unacceptable quality due to its form still being a FrankenAct, stitched together from dozens of incoherent pieces of a year-old first attempts at KS script. The art production being unfortunately but conveniently behind, I redesign and rewrite the Act 1 scenario and it's written from scratch. Despite using the same collab method as the previous one, we are able to produce a lot better script due to our improved skills, and it's the one you see in the game.

In November, Raide joins as the newest artist of the team, going through some tribulations when he tries to match his style with the rest of the artists. Blue123 starts composing some additional background music for the game as well. You can hear his work for example in the opening movie and the track that plays during Hisao's hospital scene, Rain.

The artists complete the art needed for Act 1 in spring, we assemble a release package, test and proofread it and send it out.

So in the end, it took us almost a year of learning things the hard way until the dev team set down into a (hopefully) "permanent" form that could produce something we were confident enough to release. Act 1 is not as good or bad as it is by accident. It's a product of a lot of hard work from everyone involved. We are neither prodigial geniuses or lazy tards, making a VN of KS's scale is just so hard that first timers are bound to stumble and fall down a few times. With my perfect hindsight, I now can look back, acting like I'm pro VN dev rock star and laugh at the clumsy mistakes we did, not knowing any better back then. There are so many things (almost all of them) that I'd do differently now. Anyway, not only learning to make VNs, it also took us time to learn how to act as a group of visual novel creators because this kind of free collaboration project is not a trivial work environment at all. I won't talk any more about that, except noting that we sure have learned to get along, if nothing else.

As you can see, while KS has had a two-year long development cycle so far, everything that you actually see in Act 1 was done in a bit over half a year, apart from the UI/engine code and music that carry over from older versions. While we are in better shape than ever, the future is daunting. The full game will be at least four times longer than Act 1 and at least as complex. It will be interesting to work on it. But, that story is for the future and this is where this story ends, the Wednesday of last week when Act 1 went live. That is where history ends for the time being.

So, there you are. I wrote this history mostly for myself and for the rest of the developers so that we wouldn't forget where we came from, but maybe one of you found it interesting enough to suffer through. Longwinded, full of failure, bad luck and idiotic adventures in the world of visual novels, this story has no moral, no climax and no end in sight for now, but that's how this story of ours is.