2019.3 entered the final stages of beta testing

RE: only two TECH streams per cycle. I start to wonder about the feasibility of the whole TECH streams. First TECH stream of the year will be the worst for the users as it's support ends immediately when the second one arrives and that second one has all features upcoming LTS will have. So it really boils down to having bad first version and another version that will become the good version eventually. The positive thing about this is that at least there's going to be less versions with totally different feature sets.

I still can't stop thinking that Unity could just have one TECH stream per year if they plan to dev the engine this way. This would stop people getting surprised after supposedly stable TECH release had features that got changed on next TECH release. With one TECH stream, they could just keep iteratively adding new features throughout the year and at Q3 they could just put feature freeze on it and stabilize it toward LTS.

If you think about how Unity treats current TECH streams, they really aren't meant for production so why waste effort on stabilizing the individual TECH streams during the cycle if some of that work is inevitably going to be wasted anyway.

 

Final release is only the beginning of the bug fix marathon, so expect many bug fix releases over the coming months.

 

I do think it's a good idea to limit the tech releases to two per year. Then try to get the year's LTS out by around year end, instead of half a year or so later.

I like improvements to the engine as much as anyone. But trying to get three releases in a year was obviously too ambitious.

This approach ought to lead to more stability.

 

Unfortunately, I will never ever participate in any beta / release candidate testing. This is so irritating, when not some edge cases, but the MAIN SUCCESS scenarios are broken. I lost so much time by sitting on forums, reporting bugs, relaunching Unity 20 times a day after crashes. And after all that they just pushed release button, when tons of issues was not resolved. At my previous job, this level of quality was not acceptable. Main success scenarios should work on the local verification stage even before the build reaches QA. And QA should never miss such obvious cases. I really. really don't understand what is going on and why the beta versions of 2019.3 were much more stable, than release candidates.

 

Even if they haven't fixed the bugs you've reported yet... the sooner they get into the system, the sooner they'll filter through it. They will be fixed in a patch I'm sure.

To be honest, this version of unity hasn't been the worst I've used before so it's about on par with what I expected from a beta release. Could it be better? Sure, but it is what it is.


And just an FYI, they fixed the bug I reported about post processing in URP breaking the prefab view. \o/

 

Well... 2019.3 is now live....

 

I agree 100%. Shipping with a TECH release is a Russian roulette: if no crippling bug shows up until shipping date, you're golden, otherwise you have no option other than upgrade to the next TECH which can introduce project-breaking changes as seen in 2018 (new prefab pipeline dropped in the last release) and 2019 (LWRP losing custom post process in the last release).

 

I'm still not sure what the issue is though... I mean I always thought it was:

Tech release: Expect stuff to change and break each release.

LTS: Expect no new tech but get small bugs incrementally fixed and things like iOS and Android specific compliance features to get rolled in if they popup from apple and google.

seems simple to me. If you expect to release a game, you should probably be on LTS, if you need brand spanking new tech then you may just have to wait, or you take on the burden of building a game on a moving tech base. I bet that's still faster than doing it yourself!

 

I dunno if the change made to 7.1.8 URP but there is going to be PPv2 option for it so you can technically keep using the custom PPv2 shaders while waiting for the custom PP on the URP's own internal PP to arrive.

 

Agreed. They just cannot deliver.

Three words to describe Unity accurately. Lazy, Stubborn, and Incompetent. And they decided to become lazier to cover the incompetence.

If they are going to reduce to 2 tech releases, they should put DOTS on the separate branch at least.

This is what I wanted to hear "We are reducing to 2 tech release because we decided to put DOTS on a separate branch so that it doesn't drag everyone down and decided to port the Editor to .NetCore to help everyone. Having more branches require more work, therefore we had to reduce the tech releases and I hope everyone will understand it." but instead, We are going to slow things down!

So disappointed.

 

So much grief here, hope it only comes from people who pay for Unity.

 

Lets be real here, going from 3 to 2 tech releases doesn't necessarily mean less features, but it does mean less work on stabilizing the TECH stream as they only need to do it twice now (instead of trice), so they do save tons of hours by doing this. I'm all for this part as I hate efforts that end up as waste and TECH stream releases aren't the main goal like the LTS is now.

And since nobody really should ship with TECH stream, I don't really see this decision hurting anyone. If anything it can only be better for all (at least if you start preproduction on x.1 now, you don't have to port your product more than once to get to the LTS (as x.2 is essentially going to be x.3 LTS feature and API wise).

 

Actually you cannot really compare the situation since at the moment Unity is facing a huge overhaul with all the scriptable render pipelines and DOTS. So yes, the transition is very hard and annoying for everyone but it has to be done.
Customers need better graphics and better physics and better performence.
Also, the list of asked features is very long by all users and fanbase, so I think they have too much to do, to address both bug fixes and feature sets.

Nevertheless at some point, Unity has to deliver, if they don't want to fall behind. What I saw with the beta is, that bugs were known like months and fixing them was scheduled over and over again.
For example, I was waiting despairingly for months for bug fixes like for clothes and decals. So I don't if they either don't have enough staff to do it or the internal code is too old and broken to be repaired without breaking other stuff.
Also what I see and read about anything graphic wise in Unity, I came to the conclusion that the graphic and lighting department of Unity must be really too overworked or too small in team size for the set goals. (Despite people saying Unity staff is lazy.) Work as a graphic programmer in huge complicated projects and then tell again.

And the thing with waiting for 2019.4 is, that people who need Realtime GI, cannot wait for it or use it then, because then support for Enlighten is dropped completely. Another point is, to use the better GPU lightmapper version, people are asked to stick with 2020 alpha, with is stability wise out of question. So either realtime or baked, you are completely stuck with old versions of Unity and have to get over the gap years. So for me, my dream games have to be scheluded, since the tech for the photorealism I want to achieve is simply not there or to cumbersome and old (with Unity). But at least I have the hope that Unity is now changing and in some years, will have all the features.

 

Enlighten is still in 2019.4. So far i remember, is the last supported LTS for build-in renderer und existing HDRP projects.

 

Just because you're unfortunate to work with low-quality stuff doesn't mean that everything has to be so.

 

Where was the .net core thing announced?

 

Nowhere, he stated that he wished it was announced.

 

WHOA...

I think he was referring to the overall process and past experiences with unity. Whether a product is high quality or not, the process remains the same; "You do not release something and tag it as production-ready when the number of bugs exceeds the 'acceptable' threshold amount". Although I do not know the acceptance criteria at UT, I do not think that the recent release should have happened.

 

The naming TYPO3 tech release are "sprint release", they are release steps forward the next LTS, and they recommend the LTS for production projects. Sprint release are for experience new feature, give extension developer time or for the impatient. But unlike Unity TYPO3 is not suited for hobbyists.

 

Here is some news about .Net5
https://forum.unity.com/threads/unity-c-8-support.663757/#post-5331426