Development Progress Report, week 54

This update is a couple days late, because some appointments and things got in the way.

I’ll be brief. I promise.

There’s been a mix of work going on in the last week. I’ve managed a few more optimisations for the story-compiler (which you can see in the rubbish screen-shot above), done some modularisation and speed improvements on the story-loader, and general grunt-work on the narrative.

Some of the improvements cut both the byte-count and word-count of the narrative down. I keep finding ways to compact things, which results in end-to-end improvements. Then the word-count goes up as I write new narrative. Then it goes back down again as I refine and compact things. It’s weird, and I don’t really know how long the narratives actually are, except in the most approximate sense. There’s too much dynamism involved for that. The best I could do would be to generate bunches of transcripts and base word-counts on arithmetic means.

I’ve also looked at improving the prologue. One of the characters had a whole mess of backstory in the prologue, which probably ate up 15-30 minutes before the action got rolling. That backstory has been split up and moved until later in the narrative. Most of the other characters have backstory that is doled out in pieces, but this one had hers right up-front and in your face before we could get the story moving.


One character got a little extra backstory added at the start. According to the telemetry, that one character gets picked like 90% of the time by people trying the demo out. Unfortunately, that one character also skipped over a lot of key pieces of information at the beginning, and I don’t think it helped the reader get comfortable with the setting. Also fixed.

Now, the question is why this one character keeps getting chosen by first-timers?

The only information available about each character to the first-time player is their name, and a short description. Something about Kelvin Bowyer’s name or description is significantly ‘grabbier’ than it is for the other characters. In the next week or two, I think I need to go back and at least revisit those descriptions. Either Bowyer needs to be less grabby or the other characters need to be more grabby.

Oh, and as for last week’s issue with audio lag? It was the audio buffer size! 4096 byte buffers added too much delay for S16 stereo audio at 22010hz/channel. 1024 bytes seems kind of tiny (especially if you’re someone who cares about interrupt hardware interrupt rates), but makes the audio much more responsive.

And so, on we go into the new week. Chapter two is starting to flesh out, new characters and situations have appeared, and some things have taken an unexpected turn.

See you all again, next week!