Beta Patch Notes e1.5.8

Users who are viewing this thread

While I share your sentiment, I'm not quite sure you understand how software development works (or at least QA and bugfixing). In order to fix a bug, you have to be able to reproduce it in a debug environment where you can follow the process through the code. Without a reproduction case, fixing a bug in a large project is very difficult to near impossible. Without being able to run through that reproduction case in a debug environment, fixing a bug is also very difficult to near impossible. The problem with a bug occurring in a live environment but not a test environment is that there is no way for the devs to be able to determine where in the code it is occurring, which makes it a needle in a haystack. So owning a Steam copy and testing it on there should be a part of their QA workflow, but it only helps them to be aware of problems, not to actually fix them.

I work in software development, spending a lot of my time on bug fixing, and it is very difficult without a reproduction case and being able to reproduce it in debug. I'm not saying this to excuse TW's buggy software, development process, or management, but hopefully to provide some perspective to those that don't know what it's like.
Also working in software development, there were clues that there were build issues last summer (I don't recall which betas); missing assets, bugs being hotfixed but then recurring in the next incremental release, etc. Working in software development also provides a view of how testing actually works - and it's not 'just playing the game for a few hours'; a relatively small number of testers will have test scripts focusing primarily on bug fixes and significant new features. It's impossible for TW to get anything like the depth and breadth of real player testing (and some days, there are 20,000+ players) even if they employed 100 people to just play the game, which they don't and won't.

Sometimes in software development, you have to decide between 'fix forward' or 'roll back'. Generally with an agile incremental development cycle, you fix forward, and that's how TW works with betas and hotfixes. 1.5.8 beta has been particularly buggy - and is now approaching four weeks with only one of the significant bugs having been fixed; several attempts to fix another have not succeeded or only partly succeeded; and no news at all on perhaps the most significant - the game periodically deleting/disappearing/hiding a bunch of NPCs (including two of those involved in the main story quest).

We know from various threads with TW input that 1.5.9 will already have been pretty much 'waiting to go' since 1.5.8 went into beta (a couple of betas have gone live within 2 weeks, with a new beta ready), while the developers will have already been edging towards wrapping up 1.5.10/1.6.0 and deciding on final scope of 1.5.10/1.6.0 vs stuff that would fall into 1.5.11/1.6.1. But 1.5.8 can't go live in the stable branch until those bugs are fixed. And presumably (unless the build issues are even worse than they appear), those bugs are then already present in 1.5.9 and 1.5.10/1.6.0.

By now, someone in TW must certainly have wiped their internal build and gone with a clean Steam build to replicate the issues (if they're entirely not present in internal builds; one of the bugs I'd find that difficult to credit, because it looks like pure a coding bug, not a build one). The issue should not, after 4 weeks in beta, be replication. So, that leaves us complexity of identifying the root cause(s), or complexity of fixing without breaking something else. Neither of those is good news, particularly then considering the impact is not just on 1.5.8, but on 1.5.9 and 1.5.10/1.6.0 and work that must be ready to - in theory - be the primary development focus on 1.5.11/1.6.1 by now - both code and regression testing; do those bug fixes (that we haven't had sight of yet, if TW have found any) work through three different versions of the code?

Worse, at some point any software development organisation would consider rolling back, if fixing forward isn't working. That would mean scrapping 1.5.8 and any code already in development on the same codebase - i.e. 1.5.9 and 1.5.10/1.6.0. That would be literally months of work to be (at least partially) discarded and redone.
 
It hasn't.

Which is why I'm baffled that people seriously claim horse archers are somehow accurate.
Three cycles of inhaling... Oh boi...?
giphy.gif
 
Did you ever get hit?
In that video, which I must repeat is a harmless mockery :lol: ; keeping that crouched posture I doubt very much that I would have received any arrow hits from all the arrows available for those two units.

And obviously, I didn't alter any AI behaviour or range parameters of effective fire.
 
Back
Top Bottom