Abbess Otilia's Life and Death Postmortem

As with my previous IFcomp games, I’d like to share my thoughts on the design of my entry this year. But before I get to the excuses, rationalizing and retorts, I want to express my gratitude to all those who played my game and commented on it in public or in private. I was happy about every word of encouragement I received and there were also many great ideas for improvement. Thank you all!

I find myself writing quite a lot this time, so I will split this postmortem into several posts, hopefully over the next few days. I will first post on the form, then on the content and finally on the language of Abbess Otila.Yes, I know that the language is technically part of the form. Feel free to comment in between, though. I will try to answer any questions.

1. Form

McTavish wrote that he felt Abbess Otilia was a case of an “author’s desire to produce a beautiful artifact overcoming a common sense approach to creating a usable game”. I want to admit right here that he is spot on. The game started out with me trying out different ways to improve my game Rite of Passage right after the end of the 2016 IFcomp. That game comes in the form of a diary, so I thought a book layout would be a way to improve the presentation. I set about producing a css book layout for twine and I thought what I arrived at was pretty. I quickly discovered it wasn’t so great in terms of readability, mainly because it dumps a lot of text on the screen at a time (if you want to fill the pages, that is). In fact, as some people have commented, this kind of presentation breaks some of the consensus on how to do text game interfaces. The way you are supposed to be doing it is to soft-cut in single sentences, on neutral backgrounds, in easily readable, large lettering, one after the other. Not only will this improve general readability, it gives your words so much more gravity. This way of drawing in the player is also very inclusive. It makes text games accessible to audiences that would otherwise stay clear of written material. This is important, especially from a commercial perspective, as these are, statistically, the ones most likely to pay money for your product. Honest advice: Don’t sin against the dogma.
Having done the css, I did not want it to go to waste, so I thought: what kind of game can use this without getting slaughtered at the IFcomp ballot? It was clear that it would have to be something where the layout itself carried importance, as opposed to the idea that there is a single best layout that will fit any kind of content. The conclusion I arrived at was that simulating an illuminated manuscript would be just about the only thing that stood any chance of getting away with it, if only for the fact that people would realise that trying to do this in a more reader-friendly, modern, not so terribly old-fashioned manner, would turn out even worse. I had to spend many more hours on the css. Getting it to display acceptably in all browsers and on all devices was a nightmare. I say acceptably, because I had to give up on nicely very soon. Of course, readers complained that the game looked nicer on that other guy’s system which happens to have a horizontal monitor and a browser that implements automatic hyphenation, or correctly renders drop cap initials, or fills columns evenly. The effect is - among other, perhaps lesser evils - that the resolution/browser combinations with the best rendering can’t fill the pages to capacity, as more text would spill over the page on some other machine.

One interesting aspect of medieval manuscripts is that they are, roughly speaking, easier to read the older they are. The earliest texts from the period are still written in majuscule Latin book hands, somewhat later books were written in minuscule (carolingian, isular, uncial) before blackletter scripts became dominant in the mid-to-late middle ages. Today, we mostly use Renaissance-style fonts (i.e. Garamond). These are originaly a mixture of Latin capitals and carolingian minuscule, which is why we find both roman epitaphs and the Book of Kells relatively easier to decipher than, say, a 15th-century book of hours. Depending on the specific font, blackletter can be somewhat difficult to read. The later quadrata types are especially tricky, with many letters looking almost identical. Since I was already going down this road with my choices regarding layout, I felt I couldn’t compromise on historical accuracy in my choice of font. Then again, I did want people to be able to read my game, so all that was left for me was to choose a time (and place) for my “manuscript” that would allow for an easy-to-read script. Why doesn’t the game take place in the 9th century, then? The scarcity of sources for such an early period precluded it. If I was to stand any chance of providing a reasonably realistic portrait of the customs of the time and the life of its protagonists I would have to go for a later period. The compromise I arrived at was 13th century textualis. Of course, the font we see in the game is only an approximation of what such a text would look like. There are two main reasons for this. The first has to do with the failure of unicode to encode the long s, and the second with my own choice to limit the number of ligatures and to forego abbreviations in the interest of readability. In a real 13th century manuscript, these would be a frequent occurrence, partly because they were used to manage the kind of things that are nowadays solved by justification, hyphenation and hanging punctuation.

To justify the undertaking, still more had to be done. 13th century manuscripts usually come with at least some amount of rubrication, so colourful initials for every paragraph were a natural requirement and relatively easy to implement. Chapter initials came next. I revised these several times. At first I was thinking of “alive” initials with little illustrations. These could follow the contents or just consist of one of the traditional sequences of illustrations for the months of the year. Eventually, I dropped that idea altogether, mainly because they took the reader’s attention away from the of the actual illustrations. I settled on a series of mostly ornamental initials. Since I didn’t want to re-use too many of them several times, I ended up with a number of different styles. This is not entirely unrealistic. Rubrication of a single book was often done by several artists, each in his or her own style, who split the task by chapter. My “book” is really a bit too short for this, but it provides variety I would have otherwise had to forego.
I did the illustrations only after I was finished with the writing part. They provide a way to fill pages that would otherwise have ended up half empty, and that is the most important criterion for whether there is an illumination in a particular chapter or not. They are not up to the standard I would have liked and I want to express my wholehearted respect for the illuminators of the middle ages who did this without the benefits of unlimited layers, magnification and undo. In many cases, the quality of their work is simply stellar and easily puts anything I could ever create to shame.
The game was supposed to be finished a year ago, and I had entered it as an entry for the 2017 ifcomp, but withdrew it at the last minute. When I saw Harmonia from that year, I thought: It mimics a book with illustrations and marginalia. That’s just great, now everyone will think I copied it (I very much recommend reading Harmonia, by the way). Many reviewers noted the similarities, some hinting at the possible influence with more conviction than others. The marginalia in Abbess Otilia were, however, born prior to that, simply out of my desire to give the game more of an authentic look. I therefore hadn’t included the on-click transcriptions until one of my testers told me they were too small and virtually indecipherable. I said “yes, that’s how they’d look like in a manuscript, but if you look closely, you can read them, if you want. You might need a magnifying glass. See here, this says…wait a moment, I’ll take a look at the source.” So that is where the transcription function originated. I am glad I included it and I think it is a neat feature of the hypertext format. If the marginalia didn’t provide for enough interesting content by themself, as some people complained, that is due to the fact that they started out as mere decoration.

Choice Structure
Rite of Passage had an elaborate structure of multiple branches per decision point and delayed branching, powered by a complicated system of stat variables and Boolean flags to represent the personality of the main character and the relationships between the other characters. All of this worked in the background, almost like a quality-based narrative, and it wasn’t clear to the player what was going on, therefore it felt arbitrary. While that did work towards the esthetics I had in mind for that game, I wanted to avoid the effect in my next work. It’s just not much fun to write a lot of content and mechanics that no-one will be able to appreciate. This is why Abbess Otilia has a more traditional friendly gauntlet structure, which branches out for two consecutive choices at most before it rejoins at a choke point, that being the next chapter, taking place one month on in game time from the last. The delayed branching is still there in a number of places, but I tried to make it clear to the player which prior choices led to the situation.
Other than that though, the choices are similar in nature to the ones in Rite of Passage: They concern dialogue, thoughts and higher-level decisions, as opposed to more fine-grained, physical interactions, which I feel can be done better in a parser game.

All in all, I tried to play to the strengths of the hypertext medium in both presentation and mechanics.

I absolutely loved the presentation.

[spoiler]Turning pages was great. I wish the facing pages could be a side-by-side original and text transcription.

Or…if hovering over the beautiful manuscript text would pop phrases out in a plainer font - like running a magnifying glass over the page?[/spoiler]

I’m glad you liked the page turning. I redid the page curl graphic four times from scratch until I was satisfied with it. And I actually thought about hovering as a way to transcribe the text. It would have been a neat solution, but it doesn’t work on touchscreen devices.

I don’t mean to imply you need to re-write your game, but another solution I thought of is to have the right page be blank, and have the player click or tap phrases on the calligraphy of the left page to reveal it in transcription on the right. As if the player is writing the transcription themselves. They wouldn’t need to click everything if they didn’t want to, just on occasion where it’s hard to read. Just a thought.