Dear Esther- Two Years In... (Part One)

donnelley_web

So, today is the second anniversary of the launch of Dear Esther, and we wanted to stop and take a moment to reflect. We've talked a lot about the game and its history over the last two years and we don't want to just repeat ourselves. So we thought we'd split this into two parts, firstly talking about a few things we've not really talked about before about the game, and then a bit more about where it's taken us as a studio.

1. We weren't originally going to make it. I wanted to do a take on Day of the Triffids where you had to steer a bunch of blind, panicking civilians through a dangerous environment, but we couldn't afford the AI or modelling. And didn't really know where to start.

2. The first concept art was a massive lump of clay on a desk in a disused University office. It was like Close Encounters of the Third Kind. We carved paths and bored tunnels into it whilst we tried to figure out the island design.

3. Right up until the release of the mod, jump and crouch were in the game, but we pulled them out because it felt like you got more obsessed about where you could get to rather than where you were. When Rob started on the commercial build, initially they went back in again, but then got removed again for the same reasons.

4. Rather than adding music at the end, Jessica wrote parts of the score right at the beginning of development and we used to those to inspire the island's design and the script. We still work like this, btw.

5. The original mod had one of three blocks of narrative randomly selected per cue, we added a fourth to the commercial version. This was in response to fans coming up with story interpretations we hadn't thought of which were really cool and we wanted to integrate further.

6. Nigel Carrington was the second voice over we recorded for the game. We had the entire thing originally recorded by a local actor, who was really good, but it felt wrong when we integrated it into the game. It was a really important early lesson about how insubstantial stuff like the relationship between tone of voice and tone of environment can get amplified into serious issues.

7. Parts of the landscape of the original mod were placed in code – trees were planted and paths shaped to visually represent images from the game (like the shape of the human femur, or the M5 between Wolverhampton and Exeter) even though you wouldn't be able to see them in the game. The idea was that the player would literally be walking out the story, even if they didn't know it.

8. Rob took these ideas further in the commercial version, making them more visible – the mountains in the last level make the silhouette of a sleeping woman for example.

9. Some of the music cues are coded as well – Jessica used morse code to form the rhythm patterns for a couple of tracks.

10. Rob coined the phrase “emotional signposting” to emphasize that it's not just about orientation and information, but that part of environment design is to create a framework of feeling for the player, that steers their responses to what is happening.

11. Nigel recorded the original voice-overs in a single morning. We did three takes of each cue, and ended up using the first take 90% of the time. He first read the script about 45 minutes before we started recording.

12. One of the hardest decisions we made with the commercial remake was whether or not to include head bob.

13. The narrator turning into a seagull is not because we're Brighton & Hove Albion fans, but reading it that way puts an interesting new spin on the climax of the story.

14. The only reason we took control away from the player at the end of the game was that it's pretty easy to fall off a ladder in Source and we figured even if we didn't like the idea of taking control away, it was a better option than having a player potentially fall off the ladder at the climax of the game and ruin the drama.

15. The screen staying black at the very end until the player hits a key came from literally not being able to find an amount of time for the blackout that felt right. We came to the conclusion it was better to let players decide when the time felt right. We were concerned some players would assume it had crashed (they did) but the pay-off was worth it for the majority.

16. I wrote the script for Dear Esther in about three days as a stream of consciousness. I'd been working on the island design with the original artist, Josh Short, for a couple of months, but the actual text came out pretty fast. Then there was a few weeks of balancing, trying to make sure that no one version of the story came out more strongly than any other.

17. There are some factual inaccuracies in the game, particularly about the existence (or not) of a Welcome Break service station at the Sandford junction of the M5. I apologise for this. Dear Esther was, if nothing else, a lesson in how amazing the impromptu detective skills of gamer communities are and that they really care about detail. This had a big impact on writing the script for Machine for Pigs, where I tried really hard to make sure all the detail, down to a very low level, was properly considered.

18. The first two levels of the commercial version stick very closely to the original mod's design. The caves are significantly different. The last level takes the basic design, but it's about 50-60% bigger in scale. The biggest difference in the last level is the reduction of writing on the cliff at the end. We tried many many times to make it look right, but it never really worked, so there's much less of it than the original mod. This is obviously counter-balanced by the huge amount more detail elsewhere in the game.

19. Dear Esther has been translated into fifteen languages. These translations were all made by the fan community, and that's one of the most amazing compliments a developer can get, so we are hugely grateful

20. No-one got paid for Dear Esther until release. Rob lived off ASDA pot noodles, Jess and I were both holding down other jobs. We paid our freelance coder and Nigel, but if the game hadn't sold, we'd have sunk two years into it for nothing. Doing it for the money is always a bad reason for doing it, but watching the figures on the first night of sales and realising that the risk had paid off was one of the most amazing feelings ever. We still can't quite believe how many units it sold in that first 24 hours.