Hej airport CEO and welcome to the 105th edition of the, more or less, weekly development blog! After about two weeks in the states we have finally returned back home to Sweden and am again writing to you from our office in Malmö. We have switched our part GDC, part inspiration and part vacation trip for good old regular business and Airport CEO development and since we’re coming out of a week with overall lowered activity and are currently quite jet-lagged, this post will mostly serve as a short update and a pointer towards what’s ahead for the coming week(s).
Back to Business
Since today was the first day back in the office, it has been spent dealing with the backlog that gets built up when you’re semi “off the grid” for a while, including responding to business e-mails, processing the Jira backlog and other administrative stuff such as watering the office plants (two out of three survived). The remainder of the week will be spent planning the next releases for the experimental branch as of next week which will contain bug fixes and minor feature updates as well as planning the continued work with the recently started UI sprint.
In the previous devlog we wrote about our GDC 2018 experience and one of the topics we touched upon was the large upcoming architectural changes for the Unity game engine. You can either read about it here or check out this great forum topic by forum user and moderator pderuiter. We set up a a road trip backseat lab and implemented a very basic project using the new experimental Unity version but which still would have a valid application the the ACEO use case and made some significant headway. The .gif below was captured in the Unity editor on a mid-range laptop running in power saving mode, using ECS, the new Unity job task system and Burst compiler and simulates 10000 simple agents (i.e. agents containing now transform children) at around 30 FPS. Comparing this to the around 10 FPS we get on our high-end work stations on Alpha 26 (also in Unity editor) running a similar simulation (but with more complex agents) means that these initial tests are very, very promising. The code we’ve written, as well as the underlying Unity code, are both subject to optimization but already at this stage we can see that there are large performance gains with this new approach. What’s also interesting is that it seems as if the performance impact with ECS is not exponential, as is the case with the current Alpha 26 (i.e. the game will in a multiplying manner increasingly struggle to simulate more passengers) but instead follow a more linear trajectory.
Sadly, it will be many months before we can implement and deploy this technology as part of a stable, or even experimental, version of Airport CEO since the code base provided from Unity at this stage is highly experimental. And while Unity expects ECS to become a stable and supported standard as of (estimated) Unity 2018.3, probably scheduled for release later this fall or in the beginning of 2019, what these initial tests show is that the performance gains are too big to ignore. We’ll make sure to continue these kinds of test and prepare the ACEO code base for an ECS deployment that is aligned with a stable deployment of an ECS supported version of Unity.
A Sneak Peak of the New UI
We’re steaming ahead with the development of the new UI and are now ready to reveal a first sneak peak:
April Fools’! Wow… can’t believe you fell for that one…
… okay that was a pretty bad, and late, April Fools’ “joke”. We were in the desert on the 1st of April and simply did not have enough bandwidth to execute one properly but rather late than never. Or maybe not in this case. While the development of the new UI is progressing rapidly we still have a lot of design work left to do before we’re ready to reveal our thoughts and ideas to you for feedback. What we’ve done so far is a large initial layout re-structure to account for the large technical debt contained within the UI code base. If you open either the flight planner or management panel for a busy airport in Alpha 26 you’ll notice that there’s a notable performance drop when interacting with said panel, this is the result of that technical debt which we aim to eliminate not only with a new UI layout but consequently with a new UI system architecture. We’ll share more details on this as we draw closer to a presentable product.
However, we’re currently looking for a freelance artist to help us create some new UI graphics for assets such as buttons, knobs, sliders and other components. Is that someone you? Well then don’t hesitate and drop us an e-mail and include a short presentation and a link to your portfolio with previous work too hello [ a t ] airportceo.com. We’ll be looking forward to your message!
So that’s it for this week. Thanks for keeping us company and make sure to check back with us on the 16th of April for the next edition (as there would be little sense in updating you all again on Monday). Thanks, and fly safe