I’ve tried to not write this post. I really have. But I’m finally frustrated enough with the sim and with the development direction Microsoft and Asobo are taking MSFS. It pains me to have to write a negative post about a sim that I have 1,500 hours (and hundreds of dollars invested in) but it’s gotten to that point. The main issue is the stability of the platform. In the recent Dev Q/A on Twitch, Jorg mentioned three things that really drew my attention, and convinced me that some perspective might be helpful for the team. So here goes.
- “Gaming as a service.” Jorg at one point made the comparison between FSX and MSFS and stated – correctly – that FSX was a static platform and that MSFS was built basically as an always evolving “service.” Software as a service (SaaS) in the business world makes sense when the platform is stable and the software updates are thoroughly tested before being rolled out to end users. Windows is essentially SaaS at this point, so is Microsoft Office and a wide variety of other business software.
I didn’t buy MSFS as a GaaS. I bought it as a simulator. I presumed it would be stable. I presumed that it would be a great development platform. I knew it would be updated with new content (everyone’s got to make money on this) and that there would be the occasional bug fix. But what we’ve got now is a what amounts to a new version of the sim every couple of months, which fixes some bugs, introduces new ones, and generally frustrates the end users. That leads to the second thing Jorg said:
- “We’re thinking about giving more time between updates” (paraphrasing here). I wholeheartedly agree with this, but I’m not sure the team is ready to do what I think really should be done: quarterly updates. Two world/scenery updates, and two sim/bugfix updates, per year. Separate the add-on planes from the world updates, have emergency fixes go out for major show-stopping bugs off cycle, but otherwise please commit to a cycle that will give your engineers the time they need to really need to first stabilize the platform, then fix the outlier bugs, and then finally start working on new features.
For example, Asobo is working on things like a new flight model for helicopters, adjusting the atmospheric physics engine for soaring, creating a system to more realistically represent airflow from props, etc. These are all great things and I look forward to seeing them all in the sim at some point. But you know what I’d really like to see done first? Fix the memory leak that prevents so many users – me included – from completing flights of over an hour without our frame times going to over 100 ms and our FPS going down to 2. The performance degradation bug should be an “all hands on deck” priority and take precedence over all other development efforts. Adding content to the sim isn’t going to do any good if the sim doesn’t work, and that’s the state it’s in now.
And that brings me to the last statement Jorg made:
- “Make sure you test your modules in the flighting because we can’t possibly test everything out there.” On the surface this is 100% excellent advice. But the question is…test against what? The platform isn’t stable enough to test against. Or at least, it isn’t stable long enough to be tested against thoroughly. It’s really difficult for developers to know how to do what they want to do when the rules keep changing and new features keep popping up unannounced, not to mention the new bugs that come with each new Sim Update.
I myself have wanted to take a shot at writing a program that would integrate into the sim as sort of a logbook replacement (side note: the logbook has been broken since FS2000, probably earlier. Not blaming MS/Asobo for this). But each time I start to take notes and get going on it, I realize not everything is there that I need, marked as available in the SDK when it really isn’t, etc.
Now I would never expect MS/Asobo to thoroughly test third party products. They are responsible for the platform and the add-ons they themselves have produced, nobody else’s. So yes, PMDG/Fenix/Aerosoft/FBW/Carenado/Etc. need to make sure their products work when new versions of the sim are being released. But that’s a two-way street. They also need a known starting point to test against. When PMDG delayed the 737-700 because SU9 was imminent, and they wanted to make sure it worked with it, that was a wise move. But I’d hate to have sunk all the development time and money into the -700 only to have to regression test it continually as monthly changes to the core platform keep coming out, breaking new things in their product. We all would rather have them be able to work on the next thing in their product roadmap, not keep retesting the same thing over and over again.
I’m writing this in the spirit of constructive criticism. Software engineering is hard. Having to make such a huge product work on both Xbox and PCs makes it exponentially harder. I get that. But it’s starting to feel like the team is settling into a mind set of “we’re in this for 10 years, and we’ll keep improving things over time” instead of “we’re in this for 10 years – let’s make sure the platform will continue working that entire time.” We’re two years into this 10 year timeframe, and we still have an unstable platform. A beautiful, enchanting, promising, enticing platform. But it’s also one of the most frustrating experiences I’ve ever had with any flight sim, or computer games for that matter. I’ve been using flight simulators since 1983. Hard to believe that after almost 40 years, the best looking sim out there is more frustrating that the hard-coded 2 FPS Flight Assignment: ATP where geometric cubes counted as towers, perfect pyramids were mountains, and cities were just flat polygons with random dots of lights. But honestly, ATP was more fulfilling to me than where we’re at with MSFS.
So please, MS and Asobo, I ask two things: keep working hard – we sincerely appreciate the efforts you go to in order to even give us a chance to fly in a modern sim – but also please step back and refocus on stabilizing the core platform. In the long run, I think it will make your users happier, and encourage more investment by them into the sim. I realize “new shiny” things are what continue to get people to spend money on the sim, and without income there wouldn’t be a sim, but I think you might start losing people if the core issues aren’t worked out. As I said above, the degradation problem is paramount. I know there are others. But separate core from “nice to haves” and really focus on the core issues first. Please.
Thank you for reading!