Hey there, airport CEO, and welcome to yet another edition of the airport CEO development blog! Last week we were on hold due to time constraints but this week we are back again with another update on the progression and development of Airport CEO. It’s almost as if the weekly devlog as become the biweekly devlog and while our ambition still is to churn out a development blog each week we’re constantly on the outlook for optimizing our time distribution and putting the most effort where effort is needed. Right now that effort is mainly being put into the stabilization of Alpha 25.5 which is the last update for the experimental branch and the intended Alpha 26 version for the default branch. However, more on this later and as we’ve got a lot of things to go over we should just get into it!
For little over a month now we’ve been working very hard with a sprint targeted at solving some of the most critical performance issues in Airport CEO. The sprint has targeted the most critical performance issues with the game code base and engine and when comparing the default to the current experimental branch there are, in terms of performance, some very noticeable differences when it comes to the number of frames per second each branch will provide you with. The games is reportedly running a lot better on the experimental branch, which considering additional bug fixes, also is a lot more stable than the default branch and since we’re reaching the end of this first (there will be more) performance sprint it is time to deploy the final version of Alpha 25.5 on the experimental branch and then to target a transition to Alpha 26 for the default branch.
The above issues summarize the main efforts we’ve been working on for this sprint and over half of them were deployed on the experimental branch as Alpha 25.3, the remaining issues will be deployed as of Alpha 25.5 for the experimental branch and marks the end of the first performance sprint.
The attentive CEO will notice that a core part of the new path finding system has been removed from the sprint, to be exact, the improved HPA* path finding system feature. For the last week we have with the help of a few dedicated CEOs been testing it on an internal branch and have deemed it not stable enough for default deployment. While the implementation is completed and functioning, we have a few more issues we need to investigate, mainly centered around player actions such as demolishing objects and triggering a grid update (you can track those issues here, and here). As we do not wish to further delay the default deployment of the performance updates we’ve therefor decided to withhold it. Don’t worry, the new worker thread system will be part of the next update which, even though it uses the current A-start path finding, will still be very fast and should, for any path that’s not very long and very complex, result in an instant calculation (mitigating the effect of passengers standing around and waiting and a path queue count that goes through the roof). What was clear during the internal testing was that the new system was most efficient on longer and complex paths but not on short and simpler paths, as for those paths the current A-star was faster. So we experimented with a distance slider to allow the system to either trigger either regular A-start or HPA and we think both systems can work really nice together, once we solve the remaining issues.
Even though this major sprint is drawing to a close, we will continue the hunt for extra FPS and we still have many aspects of the code that we need to go through and optimize. As we screen through our performance profiler, we constantly find new things that is in need of optimization. However, contrary to when we begun working with this sprint there is now a lot more micro-optimization to be done rather than the large scale optimizations we’ve been working with now. For example, we have identified several areas where the existing animations are taking crucial milliseconds each frame, especially costly is the moving baggage belts. For this reason we are looking into an alternative solution for animating the belts and will probably include a setting to turn the animation completely off for people that want maximum performance.
Alpha 25.5 is currently planned to be deployed on the experimental branch later this week. News on any Alpha 26 deployment, depending on the stability of Alpha 25.5, will be handled via the Forum.
Finally, with improved performance, we will begin support unlocking maps as the next update (Alpha 25.5) will include an option to unlock the previously locked map areas. With that the start world will also be reduced in favor of performance (creating, updating and executing in large node grids is performance intensive), but don’t worry, it should still be big enough for large airport constructions when all tiles are unlocked. We’ll be listening carefully to your feedback in this matter and if you discover that you need more space to fulfill your designs, please let us know! Your existing saves will keep the old size but you will be able to unlock the map at your discretion. It is still not really recommend to unlock all the tiles on the old saves due to the sheer size; such larges map naturally have an imminent risk of a more noticeable performance reduction and FPS loss.
For once we did not spend this weekend coding but instead took a trip to Sweden’s capital, Stockholm, and met with the Airport CEO artist Mitchell who was there for a short weekend vacation. While we’ve been working together for several months now, this was the first real life meeting we had and spent it with a nice lunch where we discussed many different topics ranging from the differences between the Netherlands and Sweden, stroopwafels and the weather to basically everything about Airport CEO. It was great to finally meet up and establish a foundation for the collaboration and partnership ahead!
With no time to loose we decided to kill two birds with one stone and met up with a music studio, conveniently based in Stockholm. Music has not really been a top priority for the first six months of Airport CEO since release but given the fact that we’re both huge music nerds, listening as well as playing, we really look forward to blessing Airport CEO with a great soundtrack. However, since we obviously do not have time to compose the music ourselves we decided to partner up with someone who has not only time but also great skill. Bringing in a soundtrack to Airport CEO is not only an important learning experience but also a great tool for setting the overall ambience of the game and really important to the overall quality perception of the final product. We had a great meeting at the studio where we got to check out some of the equipment and have a long and productive talk about the music that we would like to incorporate in Airport CEO. We’ll make sure to write more about this as work progresses!
As we’re finishing up for deployment of Alpha 25.5 on the experimental branch and eventually Alpha 26 on the default branch, we’re also packing our own bags and are getting ready for deployment into another part of the world. Apoapsis Studios will be attending GDC, the Game Developer Conference, in San Fransisco next week and we’re very much looking forward to meeting up with other game developers, listening to inspiring talks and learning more about the game industry and game development in general. We’re definitely new to this scene and have a lot to learn, so it’s going to be a great experience. Since the weather in Sweden is particularly horrible at this time of year we’ll be spending about two extra weeks in USA roaming the country. Don’t worry, we’re bringing computers and work will still progress, switching our surroundings as we again will revert to working with more soft components such as the new UI and implementing new assets and graphics is something we believe will be really beneficial for staying sane and boosting creativity.
Are you also heading to GDC this year and want to meet up, hang out or chat? Hit us up via Twitter and let us know!
Even though the new UI sprint has been slightly delayed due to the performance sprint obstruction, as we have spent a lot of our time on optimization and fixing performance issues, we have started with some initial drafts and plan to really kicks this into working during our stay in the US (as we will not have access to our full fleshed development rigs). We want to keep the lid on until we have something concrete, but what we can tease you with a bit is that there will be a new color scheme (of course still with that Airport CEO touch), a new layout (both for the main UI and the management panel), improved performance (no more laggy flight planner), better text scaling and Steam Workshop support. More will be revealed at a later stage and as the UI has been quite criticized (and rightfully so), we cannot stress the importance of getting this right and that we have to think very carefully of the design and implementation. We’re really looking forward to getting some good feedback from new CEOs, experienced CEOs as well as UI and UX professionals.
Nothing too large, but we just quickly wanted to mention that we’ve implemented a press kit site, updated the development page (with direct links to Jira and other information) and solved a few CSS bugs. We’ve gotten reports on that the site and Forum are very slow to load from Australia and New Zealand (as the poor sever is located in middle Sweden) are looking into implementing caching solutions to ease access for those affected.
So many awesome airport designs, thanks to everyone for those! Anyway, that’s it for this week. With the trip to GDC next week and work with deploying Alpha 25.5 and hopefully Alpha 26 we will not post a devlog, but instead bring you back to speed on the 26th of March. Thanks for reading, fly safe!
Good evening, airport CEO, or something else depending on where you are! Welcome to the one hundred and fifty first development blog in the series and another update on what we...Read more
Well hello there, airport CEO, and welcome to the one hundred and fiftieth development blog! It’s been a little over another two weeks since we last spoke and we are very, very...Read more