Hello,
Given the state of the latest patch, it seems that Asobo has difficulties testing the product. The surface area to test is gigantic and as a software engineer who has tackled complex softwares in the past, the number of tests that Asobo should pass manually for each release is easily in the thousands.
This means hiring an army of full time testers, perhaps 5 or 10 of them immediately and at least for the next couple of years. While I’ve seen that Asobo is hiring testers right now, it is unlikely they’re hiring that many testers and thus I expect more broken releases in the future.
The other solution is to rely on the community to do most of the work. You can’t beat 1 million players ! Asobo could thus run a beta version of each patch/release for 1 month before declaring it stable. I know softwares that have been working like that for years. The system works well and people are satisfied.
But that means letting people the possibility to revert to at least the previous version with the version manager. While this is not too difficult to do a priori, that also means the servers have to run the two versions in parallel. In case of catastrophy, this will save the frustrations we’ve seen with the latest patch, when thousands of players could no longer play because the patch/servers broke everything.
As a soft. eng you should be well informed of how badly opt-in reversionary realms can go. You should also be aware of how impractical it is to revert to the previous revision many days now after the publish to production. Between incompatible saves, cloud variables, and totally unsettling the players who aren’t having any issues at all… it’s not going to happen. I could see something along the lines of a canary release to allow future patches to be tested by those who opt-in and understand the instability risks of doing so.
I also have a very different opinion on this:
- Don’t revert, patch forward!
- Make small increments often instead of few big updates
Another thought is:
- Make a tiered update, don’t push it out to everyone at the same time but maybe to a group of 10,000 or let users opt in into beta updates.
This requires a stable and fast update mechanism so that people don’t get frustrated when there are many updates coming down the pipe.
I am totally down for an agile developed flight sim, of course there must be thorough testing but in the end there will be always things that slip through and only become visible when used by the broad user base out there across the world. I bet especially crashes and such things come from the myriad of drivers and hardware combinations out in the field that is hard to test on their side in the dev environment.
You forget that people with different versions can’t play on the same servers.
This is an online title. The client talks to different servers (Weather, content, marketplace, multiplayer, etc.) If they patch a multiplayer issue and only half of the people play the new version, not everyone is compatible anymore and multiplayer is broken.
So everyone must be on the same patch, unless you seperate a beta realm with their own set of beta servers. But they cant play together with the Main realm.
1 Like
That’s not really a problem, beta release players will also have the stable version anyway if they want to play with their friends.
1 Like
This is probably the best compromise to keep everyone happy with what seems to be a polarizing issue for some.
1 Like
The problem is it doesn’t solve the problem of mediocre patches that break everything and with which we are stuck once they are installed. Of course they shouldn’t exist in the first place, but given it’s now a fact of life…