[Complaint] Too many bugs remaining in 3.7.4 to consider 3.7 wrapped up. (2024)

Keeping a codebase robust and maintainable is one of most common, and overall difficult software engineering problems and most projects don't do a very good job at it. There's a reason why most software just isn't very reliable - building something that 'works' most/some of the time and building something that is fully correct are two entirely different things, and the latter is much, much more difficult than it appears on the surface, especially if it is to remain so as the rest of a codebase evolves.

And backporting changes and building patch releases isn't free, either; there's a lot of work involved to merge changes over from mainline to a patch release branch, re-test to make sure something wasn't left behind or there wasn't some unnoticed dependency on something in the newer codebase, not to mention all of whatever is going to happen for the mechanics to do a release. It's typically going to take much longer in the long run to do the same amount of work and backport it piecemeal as opposed to amortizing the process (especially if the latter alternative skips all of the backport gynmastics altogether). So it's to be expected that, however disappointing it may be as an end user, there's only a limited amount of time that's likely to be spent backporting fixes as opposed to accumulating them towards the next major release.

That being said, it's hard for me to come away from the feel that Stellaris (like, unfortunately, many - even most - software projects), has a substantial "technical debt" problem and one that, at least for the relatively short time that I've had experienced with the product, seems to be subjectively getting worse and not better over time at least according to my personal experience (for which not everyone's will be the same). That goes for both bug fixes and general usability/QoL improvements; it's ... frustrating to see how many of those are still languishing, sometimes with evidence going back to years past, for issues that I keep hitting over and over (and it's also disheartening when a new regression comes along and isn't patched out, of course).

Now, don't get me wrong, a lot of games are much, much worse than Stellaris; many game related projects, especially in the more boxed product type days, were cobbled together without much of a thought towards long term supportability or longevity, and it typically showed. Things are a bit better in some regards more recently overall now that there's more reuse of common code that's maintained by domain experts and designed for long term supportability, and you can see places where that shows in Stellaris (at least in my experience - others may have differering views based on which bugs manifest in their gameplay). For example, I've been pleasantly surprised by a lack of game client crashes in my typical use. Game engines are complex and that says to me that someone's paying attention and being ruthless (as they should!) about chasing down memory safety and other reliability problems and stamping them out before they take root and make it very difficult to claw back that level of reliability after the fact, as more changes keep raining down on the codebase meanwhile, etc..

But for the scripted, data file driven parts, especially, quality seems to be... problematic. Part of that is somewhat skewed by the nature of things; if the game crashes, it's readily apparent and a robust crash collection process can help people jump on a problem like that quickly, whereas a logic bug in some scripted event that doesn't work right might not be readily apparent, and a lot of those are bespoke one-offs that are hard to build structured testing for unless one mandates it as a part of check-in (which can seriously slow things down). Especially as a lot of the scripted logic seems to have difficulty being cleanly self-contained, with a lot of interactions that leak out to other areas, it's not surprising to me that it appears to be considerably worse off than the core engine logic.

In either case, though, I'm sure that the people working on the project aren't unaware that there are technical debt problems (that seems to be why the "Custodians" initiative was announced). But I'm certainly hoping that the priority needle skews more towards reliability & usability for future releases, as one has to stay on top of keeping the mountain of problems from growing to become too unsurmountable. That, of course, surely ends up having to be balanced with the business realities of needing to continue to fund development by releasing periodic new paid expansions given that the game isn't a subscription model etc..

</long monologue>