This is very true and just life in the middle of an engine transition. I’d go even further to some degree and say unreal (UE5) isn’t right for some projects where unity or godot might be better suited. (webgl, wasm, 2D, VR) Any engineer will evaluate the lift for said project and pulling HTML5 from 4.22 into UE5 is just an insane lift vs using another tool that does it out of the box… however as someone who has tasked themselves with going to market with “release branches” of UE3, UE4, UE5 … At the end of the day that is your “starting point” and anything you need to mod in source is your burden.
I’m simply stating relative to itself the quality bar and effort needed to be successful in this partnership (Epic Today) is a larger burden on the studio/dev than its been in the past (Epic in 8th Gen). The target platform and engine support (UE3/UE4) during the 8th gen was of a higher quality standard and required less lift than the same attempt with UE5/UE4/UE3 today. UE3 had bangers coming out well into UE4s life (Rocksteady used UE4 then went back to UE3) I’m not questioning the process or the “known” best practices for using unreal. All of that is kinda besides the context of my feedback. Its a given and should be second nature for any serious studio at this point.
This isn’t a hard thing to understand really… comparing Unreal Engine to itself over the course of Major/Minor and patch version revisions, the support, the feature stability and the expected lift from epics partners (studios) … A tool has a quality standard and accountability standard to be called “release” and failure to hold accountability to some said standard is not only bad … it’s self-destructive to the partnership. Mechanics and carpenters understand this. Engineers understand this. The vendor relationship and tool quality is very import to a final product of any said partnership and when your the customer in said relationship it is your responsibility to give said feedback.
Having absolutely no skin in the game or accountability for calling a “product” a “release” is no different than the backlash from the customer that goes directly to studios when they call their said “product” a “release” … the engine doesn’t get an out just because its the tool the studio used. That’s where our partnership and customer relationship starts and we have every right to demand accountability no differently than our customers.
When you can look at UE5’s release branches and literally not a single one can make it across the line without source modifications or heavy cherry picks/fixes from almost hundreds of PRs of which haven’t even been merged yet… and comparatively you look back and studios went to market with 4.17, 4.19, 4.22, 4.26, 4.27.2 launcher builds (keep in mind the time to get to these minor builds also vs UE5)… The writing is on the wall is what I’m saying.
You can’t be a “general purpose engine” marking yourself as “best in class” slapping “release” on your tool only for your customers to realize … dang. I gotta finish this tool because it cant do x,y,z the way it needs to yet. And be like … well that’s ok because craftsman gave me the blueprints on how to finish tempering the steel so its strong enough to torque this bolt. Some empathy and understanding is fine…
UE4 released in 2014 and by 2017 “release” branches were going all the way to production. 2022 UE5 entered preview, and we are 1 month away from the 3 year mark of release. Stability and quality should be the highest priority and that should include “game development tooling and feature stabilty” as much as UEFN and Virtual Production tooling at this point.
