I’m going to try to keep this brief, as this is intended as a brief summary for your benefit, not an expansive recap. First up, a full list of updates from the project, along with a list of tags applied to each post. There’s no hyperlinking for tags on the site, but I’m trying not to turn this into an exercise in learning Jekyll (that’s for next month!)
All Project Updates
- Continuous Delivery and Launching
- Wrapping Up Dynamic Audio
- dynamic audio
- Wrapping Up Monetisation
- Wrapping Up Cost Effective Development
- Wrapping Up Epistolary Narrative
- Wrapping Up Mixing For Location
- Wrapping Up Log Files And Analytics
- Wrapping Up Week 6: Complex Rule Systems
- Wrapping Up Week 5: Complex Rule Systems
- Wrapping Up Week 4: Mixing For Variation
- intent handlers
- adding audio
- Mixing For Variation Notes
- assembling dialogue
- Wrapping Up Week 3: Graphic Adventures In Audio
- lambda layers
- Wrapping Up Week 2: Skill Flow Builder
- skill flow builder
- Wrapping Up Week 01: Velocity
- continuous delivery
- lambda layers
And now, a short recap / follow-up on each:
Wrapping Up Velocity
The testing-first approach that started here paid dividends throughout. CLI 2.0 launched mid-project, which made
ask simulate obsolete, but otherwise, all good stuff. Lambda layers are absolutely essentially.
Possibly due to COVID-19, production of the ‘Alexa Dev Chat’ podcast has slowed or stopped. I definitely recommend the Alexa Twitch channel as an alternative.
python-statemachine will still form the basis of any game state engine going forward, on the basis of being able to fire code on state transition (and determine what verbs an individual object can support) but what it does lack is the ability to encompass flags or individual ‘atoms’ related to game state. We’ll come back to this in a second.
This situational design resource was a very late find in the project that supercedes the idea of designing by state machine. Voice skills need a less context-driven design approach around how things are handled - each interaction is more isolated than say, interacting with a traditional GUI.
Wrapping Up Skill Flow Builder
I’d like to circle back around to using Skill Flow Builder to make something in the future - probably a role playing adventure of some point. It’s still the only way to, using native capabilities, play atmospheric audio while also having audio dialog play. Also, it supports a higher audio quality for atmospheric audio than Alexa supports for hand-coded skills.