Making a game to learn Esperanto (An intense week at my school 2/2)

  • 362
  • 6
  • 2
  • English 
Apr 18, 2014 02:25
(Feel free to correct my text one chunk at a time! Thanks in advance.)

On Monday morning, the professor gave a short inaugural presentation and ended it by disclosing the compositions of the teams. Needless to say, he did well to place it at the very end of his speech, as it caused a massive uproar. Once the initial excitement had faded, the students began to slowly amalgamate accordingly.

After some arduous on-the-spot problem-solving involving chairs, tables, electric cords and aisles, I finally settled down with my team. Within minutes, ideas were already floating haphazardly at our table, and before long, we found common ground in making the game a language learning one. Soon after, we picked up on the enthusiasm of one of our team members and agreed to center the game on Esperanto. That lesser-known language would bring a fresh flavor to the game and thus allow the players to test – taste – it with an unaffected palate.

Of course, it wasn't enough, and we needed a pretext for making the player learn the language. It had to be passive, in some ways – even more so as the game was supposed to create buzz, as per the introductory speech. We thus decided that the player would assume the role of a Casanova on a mission to picking up Esperanto-speaking girls, and that he would have to bring his level on par if he was to have any chance in his endeavour. With the help of an integrated wordbook, the player would be wading his way through the dialogue, trying to understand and make himself understood. In practice, he would type his questions and answers in an input box, so that the game would have a more realistic feel and a more natural flow. We also planned on incorporating a healthy dose of humor to make the whole more engrossing.

We were all jubilant at the idea, and when the professor gave us carte blanche, we enthusiastically began work.

Unfortunately, for a long time we only had the outlines of the scenario and minigames rather than a continuous storyboard, so I didn't know where to start with the programming. On top of that, I quickly noticed that I simply wasn't productive at school; that is why I started spending more and more time developing at my place, often making up for lost time by working into the early hours of the morning. To keep myself awake, I often played some upbeat songs on my computer and exagerated my enthrallment, making wild gestures in the air and all.

As a result of that unsatisfactory workflow, we had a presentable game only on Thursday evening, while others had glossed/polished and stowed theirs as early as Tuesday. Some teams had thought up simple yet highly addictive game concepts that didn't take long to develop, such as this one: (it remains my favorite among this week's projects). It slowly dawned on us that our game was perhaps too ambitious for a four-day marathon, and I remember envying the choices of our rival teams at one point.

In our game, the player was to converse with a woman with the help of words and phrases that were listed to him. Since I obviously wasn't going to program a full-fledged AI, I resorted to putting stock in the user's common sense. For example, after the woman asks for the player's name and lets out a chirpy "Pleased to meet you, {name}", he would need to reciprocate by saying "What's your name?", or "What's yours?", or "How about you?" in Esperanto if he wanted to get on with the conversation. By the same token, I expected the user to respond to thanks with "You're welcome", and so on.

As mentioned earlier, another aspect of the game that we had all agreed on was its humorous, offbeat side. That translated ingame by the player's every line being enunciated in a robotic voice using Google Translate, while the woman's sentences were said out loud using pre-recorded audio files I had found on the Internet. Of course, I had limited resource, and each sentence of the woman had to come straight from the phrasebook. That limited considerably the field of possibilities.

However, those technical limitations also called for inventive ways to circumvent them, or ways to acknowledge them by using self-derision for humorous purposes. For example, when the woman said "Pleased to meet you, {name}", the first part was said in a natural voice, while the name of the player was pronounced by Google Translate. She was, in a way, pronouncing his name exactly as she had heard it a moment before. Likewise / In the same way, one audio file I had downloaded went "Mia nomo estas" [My name is], but it stopped right there. I turned the problem upside down by isolating two syllables from another recording and grafting them into the original file. In the end, the composite audio file said "Mia nomo estas Kinde" – "Kin" and "de" being the first two syllables that had fallen into my hands.

Despite that clumsy implementation of text-to-speech, I often caught myself relying on my auditory memory to compose my sentences when I tested the game. For instance, I remember being able to spell "Kio estas via nomo ?" ("What's your name?") upon playing back in my head the recording of the woman asking my name moments before. To this day, I still clearly remember the timbre and inflections of her voice; way to go to learn to speak a language!

You might think that we went through undue trouble, and that we could have gotten over the phrasebook limitation by posting our script on Lang-8 to have it corrected, and then on Rhinospike to have it read aloud and recorded by fluent speakers. We considered the idea but decided it would take too long and be too much of a hastle. What's more, diving headlong into natural conversations could easily stun a first-timer.

So, I presented my work to my team on Thursday evening. They were all grabbed, save for one person, who insisted – although empathically – that the game was way too counter-intuitive and laborious, and that it was no fun constantly referring to a phrasebook to have a chance at progressing in the game. Most people who tested the game seemed to feel the same way. I was a bit beat down, knowing that we would turn in a work that didn't do justice to what we had in mind, to what we were aiming for. We had to upload the game at nine in the morning the next day at the latest, and we also had to prepare our presentation in front of the professor, scheduled for quarter to eleven that same following day. That gave us very little time to rework our game. One of us floated the idea of shifting our target to young children. That was out of the question for me, as it meant outright creating a new game / recasting our game and changing the one thing we had all agreed on since the beginning. We discoursed understandingly until 9 in the evening at school, and, eventually, we all agreed that our game should be more centered on guided mini-games rather than fend-for-yourself conversations. And, since none of us wanted to throw all of our work out the window, we managed to find a compromise between the two versions of the game: there would still be humorous dialogue, but it would be in French. That led us to deciding on a new scenario, in which the player pretends to master Esperanto in order to impress a French girl. Unfortunately, that meant the end of our text-to-speech trip that had seemed so effective to me. The game had lost one of its biggest assets.

By the time I came back home and ate dinner, it was already past ten.

I turned on the French eclectic radio station "FIP" and resignedly put myself to work, knowing I was in for a long night. Since I wasn't limited with pre-composed phrases, I managed to come up with the dialogue on the spot without too much trouble. Thanks to all the prior groundwork I had made as regards the structure of the code, it was also easier for me to adapt to what had been agreed upon. It didn't change the fact that I went to bed at four in the morning, after listening to FIP for four hours and a half non stop. Of course, I botched and oversaw a couple of things at the end (objects concealing one another, bugs related to drag-and-drop, and so on), but I couldn't be asked any longer. I was exhausted but at the same time hugely content and grateful for the salutory feedback of my colleague, as the end result was much better than the original version. As it turned out, heeding my peers' opinions had widened my field of view and elicited the borders of the glass pane I was apparently so impatient to smash into.

The next day, I listed what I would talk about in my part of the presentation on my way to school, and when it came time to present our game, I pretty much winged it, occasionally throwing in some technical jargon to appear more knowledgeable.

In the end, everything went well. My teammates did their parts magnificiently, and – even though it hardly met the "virality" criterion in my eyes – the game thankfully made a good impression on the jury.

Here's the game in French in case you're curious: . Click "Ludas" and then "Commencer" to start the game; you will need to drag objects and drop them over the woman at some point.

We were ranked third out of thirteen teams, and we were tied for the first place for the game itself. Considering how straining the week had been and how anxious we had been feeling until the very last minute, our stepping off the stage was onto a land of relief – marking the beginning of vacation at the same time.

We reveled in a warm group hug and lived happily ever after.

Thanks for correcting!