The above screenshot is from the very first town in the next Shining Online demo. Like other Shining Force games, you start alone and have to recruit a team before you can venture into the wide open world.
All of the above is handled by a little bit of lispy-script:
Shining Online uses a custom lisp engine for scripting, and has an entity + component system for managing in-game objects. Entities can be anything from characters walking around to hidden triggers or spawn zones.
So far this setup is working really well. Entities can have scripts attached to them, and it’s possible to add components to entities via script. I’m not 100% happy with how inspecting works; it’s far too complicated for what it is, but it works so I’m trying to leave it alone.
I’m currently working through milestone 3 – items. At this point it’s mostly UI work that is left. It’s not one of my strong points, but it’s slowly getting there. My aim is to have it all complete by the end of the month so that I can start on shops and the priest.
The main screen for managing items looks like this
I’m not totally happy with the tooltip yet, but it’s pretty close to what I’m aiming for. Giving items to other characters works, but the table needs updating with stat changes. Equipping items and armor is getting a slight revamp, but should be working by the time March rolls around.
I’m looking forward to working on something a little more fun.
Here’s some boring words about what’s been going on with Shining Online since the last worklog entry (April 2018). I’m going to stick in a bunch of screenshots today to try and make it less wordy.
But don’t worry, there’s plenty of technical words (and screens) right at the end!
This is something the original Shining Online never had, but I really wanted them in the game.
The first attempt was not a total success.
After adding some pauses and timing tweaks, the finished result looks like this:
I’m pretty happy with how it turned out.
Custom text boxes
Text boxes can have custom frames, backgrounds and fonts. They can also be positioned anywhere on the screen.
Not very exciting, but at least we know what’s in the tent.
Towns aren’t the only thing that have had some attention.
Battle ranges are a key part of Shining-style battles, and now they’re in Shining Online. They take land types and movement types into account, so flying entities will ignore trees, whereas centaurs will be slowed down by them.
Finding items and gold
What kind of RPG would this be if you can’t raid people’s homes for their hard-earned cash?
Items work in a similar way, and are automatically passed to the next free character. If there’s no space, it will be left in the container until some room is available.
There’s a fair amount of behind-the-scenes work going on in that screenshot, most of it handled via script:
The text for “Cody looking in the barrel” looks like this in the configuration file “%CHARACTER% looked in the %CONTAINER%”. The tokens are replaced by fetching data from the team (the team leader’s name) and the entity being searched (the barrel).
Gold is added to the player’s gold reserves
Flags are set in the session to make sure searching the same place doesn’t give rewards twice.
There’s some other bits going on, like pausing player input, displaying a text box and playing a jingle (which you can’t hear in the gif, but it’s there).
Items are slightly more complicated; the game has to check there is at least one character with an open inventory slot, and then has to either pass it to them or display a “There is no room” message.
Here’s how the barrel in the screenshot is configured:
All entities in Shining Online are made up of components. Entities are created using a simple template system, which sets out which components are required and what their field values are. The base_money_container template looks like this:
It listens for an inspectevent, and then disables input, runs the gold:findscript using properties from the entity and then re-enables the entity when done.
The %self% variable is always set to the entity’s identifier. So for the money container in the example, it will be the string “vagabond_gold_container_1”. Flag names can be overridden by setting a flag_name property on the entity, but that’s optional.
And finally, here’s what the gold:find function looks like:
For comparison, this is what the item:find function looks like:
It’s a little more complicated, but it’s all built and configured in a similar way to the gold container.
That’s all (for now)
I’d like to get item management finished in the next few months so I can move on to the battle system. Pathfinding and movement ranges are all working, so with some AI and proper turn management there will be an actual game soon!
The last month or so have been spent on building the user interface and making it actually usable. Equipping weapons isn’t exactly a thrilling part of the Shining experience, but it’s an important one.
Right now Shining Online’s item screen looks something like this:
Even though it’s a simple screen there’s quite a lot going on.
The equipment and inventory items are entirely data driven (finally). That means the text, icons, backgrounds and slots aren’t hard-coded anywhere. The Silver Ring icon is ugly as sin, but I’m pretty happy with the others.
In the screenshot Cody is on his lonesome, but there are other characters in town that will join up and end up in that window. Character stats are available by moving right on the table, just like in other Shining Force games. Naturally it’s also data driven.
Even the window layout is defined in a file rather than in code. Coupled with some dev tools this makes it possible to tweak appearances and see updates in real-time.
The table at the bottom was extremely time-consuming and has given me a new appreciation for how browsers display HTML tables. The current table implementation supports headers, variable column widths, column alignments and smooth scrolling.
So far I’ve added the ability to equip and un-equip weapons and accessories. Dropping items shouldn’t take too long to add, but I’m expecting giving items to team mates to be a bigger job.
It seems like every project update usually contains a bunch of notes about boring engine internals. So here’s a screenshot of Demo 5’s first battle instead.
Some of the graphics are still placeholders, but it’s slowly starting to look like a game. And it wouldn’t be the same game without a bushman to ruin your day.
One thing I really wanted to improve on from previous demos was how the battlefield affects movement. There are already some slow and defensive paths, as well as some aggressive but risky ones.
Internally there have been some big improvements made. Basic character inventory and magic screens are up and running, although they’re rather ugly and will be getting a redesign at some point.
I also spent some time building tools to help with development. Seeing as I spent so much time trying to make everything data-driven, it seemed silly to have to restart the game to load new changes. With that in mind, I built three tools:
blamit – Listens for changes to a directory, and when a change occurs will execute one or more build targets. This is used to listen for changes to the various data folders, and then to re-pack everything and copy it to the project directory.
listener – Starts a server and listens for changes to a directory, and sends a message to every connected client when a change is made. The main game connects to this server during startup and will reload resources when a message is sent. If an interface screen is showing (such as the character summary screen) it will reload its definition file and redraw itself. This makes it much quicker to tweak screen layouts.
slime-server – This is built into the engine as a service. When running, slime clients can connect to the game and run script commands. This makes it easier to spawn entities or move them around without recompiling or reloading things.
Even though they took a while to develop, they’ve made development go a little quicker. But more importantly, they’ve made some of the boring parts much more pleasant to create.
The last couple of months have been a bit quiet on the development front, but this year I really want to release the elusive Demo 5. There’s still a lot of pieces that need to fall into place, but every hour spent on the project brings it a little bit closer to release. At least that’s the theory.
The biggest feature I’ve been working on lately has been the status menu framework. It’s one of those things that seemed simple on paper but ended up taking far too long.
Screen stacking was definitely the biggest sticking point. Players need to be able to navigate to new menu screens, but also need to return to the previous screen in the state they left it. Sprinkle in some interactive panels and transition animations and things start to get complex. It’s undergone two rewrites since I started, but I’m satisfied with how it works now.
At the moment the team list and character details screen are implemented, but I’m planning to have the rest finished soon™ so I can move onto something more enjoyable.